// Inlined by build_pages.py from site_data.json.
ns = ({"groups":["Peptide recall","Peptide precision","Amino-acid precision","Amino-acid recall","Precision against coverage","Area under that curve","FDR","PTM precision or recall","Calibration","Mass-based","Sequence similarity"],"papers":[{"id":1,"slug":"de-novo-peptide-sequencing-with-instanovo-accurate-database-free-peptide","title":"De novo peptide sequencing with InstaNovo: Accurate, database-free peptide identification for large scale proteomics experiments","methods":"InstaNovo, InstaNovo+","year":2023,"read":true,"reports":["Peptide recall","Peptide precision","Amino-acid precision","Amino-acid recall","Precision against coverage","Area under that curve","FDR"],"rule":"0.1 / 0.5 Da mass rule","il":"Equal","naming":null,"new_metrics":[],"follows":null},{"id":2,"slug":"instanovo-enables-diffusion-powered-de-novo-peptide-sequencing-in-large-scale","title":"InstaNovo enables diffusion-powered de novo peptide sequencing in large-scale proteomics experiments","methods":"InstaNovo, InstaNovo+","year":2025,"read":true,"reports":["Peptide recall","Peptide precision","Amino-acid precision","Amino-acid recall","Precision against coverage","FDR"],"rule":"Not stated","il":"Not stated","naming":null,"new_metrics":[],"follows":"ref. 49 (Beslic et al. 2023, Brief. Bioinform.) formulation, per 'We formulate our metrics as done in ref. 49'"},{"id":3,"slug":"instanovo-p-a-de-novo-peptide-sequencing-model-for-phosphoproteomics-preprint","title":"InstaNovo-P: A de novo peptide sequencing model for phosphoproteomics","methods":"InstaNovo-P","year":2025,"read":true,"reports":["Peptide recall","Peptide precision","Area under that curve","FDR"],"rule":"Exact sequence","il":"Equal","naming":null,"new_metrics":[],"follows":"InstaNovo (\"For further details, see the InstaNovo24 paper.\")"},{"id":4,"slug":"pairwise-attention-leveraging-mass-differences-to-enhance-de-novo-preprint","title":"Pairwise Attention: Leveraging Mass Differences to Enhance De Novo Sequencing of Mass Spectra","methods":"Pairwise","year":2025,"read":true,"reports":["Peptide recall","Precision against coverage"],"rule":"0.1 / 0.5 Da mass rule","il":"Not stated","naming":"precision over all spectra","new_metrics":[],"follows":"DeepNovo matching methodology; Casanovo peptide score"},{"id":5,"slug":"dinovo-high-coverage-high-confidence-de-novo-peptide-sequencing-using-mirror","title":"DiNovo: high-coverage, high-confidence de novo peptide sequencing using mirror proteases and deep learning","methods":"DiNovo","year":2025,"read":true,"reports":["FDR"],"rule":"Not stated","il":"Not stated","naming":null,"new_metrics":["TD mapping FDR"],"follows":null},{"id":6,"slug":"pepgo-a-deep-learning-and-tree-search-based-model-for-de-novo-peptide-sequencing","title":"PepGo: a deep learning and tree search-based model for de novo peptide sequencing","methods":"PepGo","year":2025,"read":true,"reports":["Peptide recall","Amino-acid recall"],"rule":"Exact sequence","il":"Not stated","naming":null,"new_metrics":[],"follows":null},{"id":7,"slug":"regressor-guided-diffusion-model-for-de-novo-peptide-sequencing-with-explicit","title":"Regressor-guided Diffusion Model for De Novo Peptide Sequencing with Explicit Mass Control","methods":"DiffuNovo","year":2026,"read":true,"reports":["Peptide recall","Amino-acid precision","Area under that curve","PTM precision or recall","Mass-based"],"rule":"Not stated","il":"Not stated","naming":null,"new_metrics":["Plausible Prediction Rate","Mass Error (MAE)"],"follows":"NovoBench (\"All experimental settings in this paper adhere to the NovoBench benchmark\")"},{"id":8,"slug":"accurate-de-novo-sequencing-of-the-modified-proteome-with-omninovo","title":"Accurate de novo sequencing of the modified proteome with OmniNovo","methods":"OmniNovo","year":2025,"read":true,"reports":["Peptide recall","Precision against coverage","FDR","PTM precision or recall"],"rule":"Exact sequence","il":"Equal","naming":null,"new_metrics":["Hyperscore-based proxy FDR (database calibration)","Delta Hyperscore","Backbone Accuracy / Modification Site Accuracy"],"follows":"Casanovo (residue-level precision protocol); MSFragger hyperscore; Wen et al. entrapment"},{"id":9,"slug":"bidirectional-representations-augmented-autoregressive-biological-preprint","title":"Bidirectional Representations Augmented Autoregressive Biological Sequence Generation","methods":"CrossNovo","year":2025,"read":true,"reports":["Peptide recall","Amino-acid precision","Precision against coverage","PTM precision or recall"],"rule":"0.1 / 0.5 Da mass rule","il":"Not stated","naming":"precision over all spectra","new_metrics":[],"follows":null},{"id":10,"slug":"modanovo-a-unified-model-for-post-translational-modification-aware-de-preprint","title":"Modanovo: A Unified Model for Post-Translational Modification-Aware de Novo Sequencing Using Experimental Spectra from In Vivo and Synthetic Peptides","methods":"Modanovo","year":2025,"read":true,"reports":["Peptide recall","Precision against coverage","Area under that curve"],"rule":"Exact sequence","il":"Equal","naming":null,"new_metrics":[],"follows":null},{"id":11,"slug":"advancing-proteomic-analyses-with-graph-based-deep-learning-protein-inference","title":"Advancing Proteomic Analyses with Graph-Based Deep Learning: Protein Inference and DIA De Novo Peptide Sequencing","methods":"DIANovo","year":2025,"read":true,"reports":["Peptide recall","Amino-acid precision","Amino-acid recall","FDR"],"rule":"Exact sequence","il":"Not stated","naming":null,"new_metrics":[],"follows":null},{"id":12,"slug":"improvements-to-casanovo-a-deep-learning-de-novo-peptide-sequencer-preprint","title":"Improvements to CasaNovo, a deep learning de novo peptide sequencer","methods":"Casanovo","year":2025,"read":true,"reports":["Precision against coverage","FDR","Calibration"],"rule":"Not stated","il":"Equal","naming":null,"new_metrics":[],"follows":null},{"id":13,"slug":"diffnovo-a-transformer-diffusion-model-for-de-novo-peptide-sequencing","title":"DiffNovo: A Transformer-Diffusion Model for De Novo Peptide Sequencing","methods":"DiffNovo","year":2025,"read":true,"reports":["Peptide precision","Amino-acid precision","Amino-acid recall"],"rule":"0.1 / 0.5 Da mass rule","il":"Not stated","naming":null,"new_metrics":[],"follows":null},{"id":14,"slug":"zero-shot-de-novo-peptide-sequencing-with-open-post-translational-modification","title":"Zero-Shot De Novo Peptide Sequencing with Open Post-Translational Modification Discovery","methods":"RNovA","year":2025,"read":true,"reports":["Peptide recall","Amino-acid precision","Amino-acid recall","FDR","PTM precision or recall"],"rule":"0.1 / 0.5 Da mass rule","il":"Not stated","naming":null,"new_metrics":["15N-isotopic corrected de novo error rate","PTM Site Identification Rate"],"follows":"GraphNovo (\"we adopted the evaluation metrics originally defined in the GraphNovo study\"); DeepNovo/PointNovo matching definition"},{"id":15,"slug":"massnet-billion-scale-ai-friendly-mass-spectral-corpus-enables-robust-de-novo","title":"MassNet: billion-scale AI-friendly mass spectral corpus enables robust de novo peptide sequencing","methods":"XuanjiNovo","year":2025,"read":true,"reports":["Peptide recall"],"rule":"Not stated","il":"Not stated","naming":null,"new_metrics":[],"follows":null},{"id":16,"slug":"curriculum-learning-for-biological-sequence-prediction-the-case-of-de-preprint","title":"Curriculum Learning for Biological Sequence Prediction: The Case of De Novo Peptide Sequencing","methods":"RefineNovo","year":2025,"read":true,"reports":["Peptide recall","Amino-acid precision"],"rule":"Not stated","il":"Not stated","naming":null,"new_metrics":[],"follows":"Casanovo (Yilmaz et al., 2022) matching criteria"},{"id":17,"slug":"latent-imputation-before-prediction-a-new-computational-paradigm-for-preprint","title":"Latent Imputation before Prediction: A New Computational Paradigm for De Novo Peptide Sequencing","methods":"LIPNovo","year":2025,"read":true,"reports":["Peptide precision","Amino-acid precision","Amino-acid recall","Area under that curve","PTM precision or recall"],"rule":"Not stated","il":"Not stated","naming":null,"new_metrics":[],"follows":"NovoBench (baseline results provided by NovoBench; metric names match NovoBench)"},{"id":18,"slug":"a-transformer-based-semi-autoregressive-framework-for-high-speed-and-accurate","title":"A transformer-based semi-autoregressive framework for high-speed and accurate de novo peptide sequencing","methods":"TSARseqNovo","year":2025,"read":true,"reports":["Peptide precision","Amino-acid precision","Amino-acid recall"],"rule":"Not stated","il":"Not stated","naming":null,"new_metrics":[],"follows":"DeepNovo / PointNovo / CasaNovo standards (stated generically)"},{"id":19,"slug":"xa-novo-an-accurate-and-high-throughput-mass-spectrometry-based-de-novo","title":"XA-Novo: an accurate and high-throughput mass spectrometry-based de novo sequencing technology for monoclonal antibodies and antibody mixtures","methods":"XA-Novo","year":2026,"read":true,"reports":["Peptide recall","Amino-acid precision","Amino-acid recall"],"rule":"Not stated","il":"Distinct","naming":null,"new_metrics":[],"follows":null},{"id":20,"slug":"distilling-non-autoregressive-model-knowledge-for-autoregressive-de-novo","title":"Distilling Non-Autoregressive Model Knowledge for Autoregressive De Novo Peptide Sequencing","methods":"CrossNovo","year":2025,"read":true,"reports":["Peptide recall","Amino-acid precision","Precision against coverage"],"rule":"0.1 / 0.5 Da mass rule","il":"Not stated","naming":"precision over all spectra","new_metrics":[],"follows":"'we followed all previous work' (generic); baselines Casanovo V1/V2, ContraNovo, PrimeNovo"},{"id":21,"slug":"pi-primenovo-an-accurate-and-efficient-non-autoregressive-deep-learning-model","title":"π-PrimeNovo: an accurate and efficient non-autoregressive deep learning model for de novo peptide sequencing","methods":"π-PrimeNovo","year":2025,"read":true,"reports":["Peptide recall","Amino-acid precision","Amino-acid recall","Precision against coverage","PTM precision or recall"],"rule":"Not stated","il":"Not stated","naming":null,"new_metrics":[],"follows":null},{"id":22,"slug":"renovo-retrieval-based-de-novo-mass-spectrometry-peptide-sequencing","title":"ReNovo: Retrieval-Based De Novo Mass Spectrometry Peptide Sequencing","methods":"ReNovo","year":2024,"read":true,"reports":["Peptide recall","Amino-acid precision","Amino-acid recall","Precision against coverage","Area under that curve"],"rule":"0.1 / 0.5 Da mass rule","il":"Not stated","naming":"precision over all spectra","new_metrics":[],"follows":"NovoBench-style metric set (Seven-species, Nine-species, HC-PT); not explicitly stated"},{"id":23,"slug":"disentangling-the-complex-multiplexed-dia-spectra-in-de-novo-peptide-sequencing","title":"Disentangling the Complex Multiplexed DIA Spectra in De Novo Peptide Sequencing","methods":"DIANovo","year":2024,"read":true,"reports":["Peptide recall","Amino-acid precision","Amino-acid recall"],"rule":"Exact sequence","il":"Not stated","naming":null,"new_metrics":[],"follows":null},{"id":24,"slug":"ranknovo-a-universal-reranking-approach-for-robust-de-novo-peptide-sequencing","title":"RankNovo: A Universal Reranking Approach for Robust De Novo Peptide Sequencing","methods":"ByNovo, RankNovo","year":2024,"read":true,"reports":["Peptide recall","Amino-acid precision"],"rule":"0.1 / 0.5 Da mass rule","il":"Not stated","naming":null,"new_metrics":["PMD (Peptide Mass Deviation)","RMD (Residual Mass Deviation)"],"follows":"Casanovo (Yilmaz et al., 2022) evaluation of 9-species-V1"},{"id":25,"slug":"transforming-de-novo-peptide-sequencing-by-explainable-ai","title":"Transforming de novo peptide sequencing by explainable AI","methods":"π-xNovo","year":2024,"read":true,"reports":["Peptide recall","Amino-acid precision","Amino-acid recall"],"rule":"Not stated","il":"Not stated","naming":null,"new_metrics":[],"follows":null},{"id":26,"slug":"sequence-to-sequence-translation-from-mass-spectra-to-peptides-with-a","title":"Sequence-to-sequence translation from mass spectra to peptides with a transformer model","methods":"Casanovo","year":2024,"read":true,"reports":["Peptide precision","Amino-acid precision","Precision against coverage","Area under that curve","FDR"],"rule":"0.1 / 0.5 Da mass rule","il":"Not stated","naming":null,"new_metrics":[],"follows":"DeepNovo (Tran et al.) matching definition"},{"id":27,"slug":"transnovo","title":"TransNovo","methods":"TransNovo","year":2024,"read":false,"reports":[],"rule":null,"il":null,"naming":null,"new_metrics":[],"follows":null},{"id":28,"slug":"powernovo-de-novo-peptide-sequencing-via-tandem-mass-spectrometry-using-an","title":"PowerNovo: de novo peptide sequencing via tandem mass spectrometry using an ensemble of transformer and BERT models","methods":"PowerNovo","year":2024,"read":true,"reports":["Peptide recall","Amino-acid precision","Precision against coverage","PTM precision or recall"],"rule":"Not stated","il":"Not stated","naming":null,"new_metrics":[],"follows":"Beslic et al. antibody benchmark settings reproduced (Experiment 2)"},{"id":29,"slug":"a-transformer-model-for-de-novo-sequencing-of-data-independent-acquisition-mass","title":"A transformer model for de novo sequencing of data independent acquisition mass spectrometry data","methods":"Cascadia","year":2025,"read":true,"reports":["Precision against coverage","FDR","Calibration"],"rule":"Not stated","il":"Not stated","naming":null,"new_metrics":["Proteome-match precision–coverage for DIA (distinct peptides matching reference proteome)"],"follows":"Casanovo decoding; DeepNovo-DIA precursor-level evaluation for the Plasma comparison"},{"id":30,"slug":"transformer-based-de-novo-peptide-sequencing-for-data-independent-postprint","title":"Transformer-Based De Novo Peptide Sequencing for Data-Independent Acquisition Mass Spectrometry","methods":"DiaTrans","year":2024,"read":true,"reports":["Peptide recall","Amino-acid precision","Amino-acid recall","Precision against coverage","Area under that curve"],"rule":"0.1 / 0.5 Da mass rule","il":"Not stated","naming":"precision over all spectra","new_metrics":[],"follows":"metrics 'consistent with those utilized in DeepNovo-DIA and Casanovo'"},{"id":32,"slug":"adanovo-adaptive-de-novo-peptide-sequencing-with-conditional-mutual-information","title":"AdaNovo: Adaptive De Novo Peptide Sequencing with Conditional Mutual Information","methods":"AdaNovo","year":2024,"read":true,"reports":["Peptide precision","Amino-acid precision","Precision against coverage","Area under that curve","PTM precision or recall"],"rule":"0.1 / 0.5 Da mass rule","il":"Not stated","naming":null,"new_metrics":["PTM precision (amino acids with PTMs identification precision)"],"follows":"DeepNovo definitions (Tran et al. 2017); text mirrors Casanovo (Yilmaz et al. 2022) wording; Casanovo reproduced with original settings"},{"id":33,"slug":"bidirectional-de-novo-peptide-sequencing-using-a-transformer-model","title":"Bidirectional de novo peptide sequencing using a transformer model","methods":"NovoB","year":2024,"read":true,"reports":["Peptide recall","Amino-acid precision","Precision against coverage"],"rule":"0.1 / 0.5 Da mass rule","il":"Not stated","naming":null,"new_metrics":[],"follows":"PointNovo/Casanovo evaluation metric (stated)"},{"id":34,"slug":"introducing-pi-helixnovo-for-practical-large-scale-de-novo-peptide-sequencing","title":"Introducing π-HelixNovo for practical large-scale de novo peptide sequencing","methods":"π-HelixNovo","year":2024,"read":true,"reports":["Peptide recall","Amino-acid precision","Amino-acid recall"],"rule":"0.1 / 0.5 Da mass rule","il":"Not stated","naming":"recall over predictions","new_metrics":[],"follows":"\"the same evaluation metrics as other de novo sequencing models\" (no specific code named)"},{"id":37,"slug":"deep-learning-driven-fragment-ion-series-classification-enables-highly-precise","title":"Deep learning-driven fragment ion series classification enables highly precise and sensitive de novo peptide sequencing","methods":"Spectralis","year":2024,"read":true,"reports":["Sequence similarity"],"rule":"Not stated","il":"Equal","naming":null,"new_metrics":["Change precision / change recall (bin reclassification)"],"follows":null},{"id":38,"slug":"contranovo-a-contrastive-learning-approach-to-enhance-de-novo-peptide-sequencing","title":"ContraNovo: A Contrastive Learning Approach to Enhance De Novo Peptide Sequencing","methods":"ContraNovo","year":2024,"read":true,"reports":["Peptide recall","Amino-acid precision","Precision against coverage"],"rule":"0.1 / 0.5 Da mass rule","il":"Not stated","naming":"precision over all spectra","new_metrics":[],"follows":"Casanovo (Yilmaz et al. 2022) evaluation metrics (stated)"},{"id":39,"slug":"accurate-de-novo-peptide-sequencing-using-fully-convolutional-neural-networks","title":"Accurate de novo peptide sequencing using fully convolutional neural networks","methods":"PepNet","year":2023,"read":true,"reports":["Peptide recall","Amino-acid precision","Precision against coverage"],"rule":"Exact sequence","il":"Equal","naming":null,"new_metrics":[],"follows":null},{"id":41,"slug":"mitigating-the-missing-fragmentation-problem-in-de-novo-peptide-sequencing-with","title":"Mitigating the missing-fragmentation problem in de novo peptide sequencing with a two-stage graph-based deep learning model","methods":"GraphNovo","year":2023,"read":true,"reports":["Peptide recall","Amino-acid precision","Amino-acid recall"],"rule":"Not stated","il":"Not stated","naming":null,"new_metrics":["Node recall and node precision"],"follows":"DeepNovo and PointNovo metrics (stated)"},{"id":42,"slug":"seqnovo-de-novo-peptide-sequencing-prediction-in-iomt-via-seq2seq","title":"SeqNovo: De Novo Peptide Sequencing Prediction in IoMT via Seq2Seq","methods":"SeqNovo","year":2023,"read":true,"reports":["Peptide recall"],"rule":"Not stated","il":"Not stated","naming":null,"new_metrics":[],"follows":null},{"id":44,"slug":"dpnovo-a-deep-learning-model-combined-with-dynamic-programming-for-de-novo","title":"DpNovo: A DEEP LEARNING MODEL COMBINED WITH DYNAMIC PROGRAMMING FOR DE NOVO PEPTIDE SEQUENCING","methods":"DpNovo","year":2023,"read":true,"reports":["Peptide recall","Amino-acid precision","Amino-acid recall"],"rule":"0.1 / 0.5 Da mass rule","il":"Not stated","naming":null,"new_metrics":[],"follows":"DeepNovo metrics (stated)"},{"id":45,"slug":"biatnovo-an-attention-based-bidirectional-de-novo-sequencing-framework-for-data","title":"BiATNovo: An Attention-based Bidirectional De Novo Sequencing Framework for Data-Independent-Acquisition Mass Spectrometry","methods":"BiATNovo","year":2024,"read":true,"reports":["Peptide recall","Amino-acid precision","Amino-acid recall","Precision against coverage"],"rule":"Not stated","il":"Not stated","naming":null,"new_metrics":["difference-i"],"follows":"\"evaluation metrics consistent with prior research\" (none named)"},{"id":46,"slug":"pgpointnovo-an-efficient-neural-network-based-tool-for-parallel-de-novo-peptide","title":"PGPointNovo: an efficient neural network-based tool for parallel de novo peptide sequencing","methods":"PGPointNovo","year":2023,"read":true,"reports":[],"rule":"Not stated","il":"Not stated","naming":null,"new_metrics":[],"follows":"PointNovo (benchmarks and parameters reused)"},{"id":47,"slug":"paser-novor-real-time-de-novo-sequencing-for-4d-proteomics-applications","title":"PaSER Novor: Real-time de novo sequencing for 4D-Proteomics applications","methods":"PaSER Novor","year":2023,"read":false,"reports":[],"rule":null,"il":null,"naming":null,"new_metrics":[],"follows":null},{"id":48,"slug":"denovo-gcn-de-novo-peptide-sequencing-by-graph-convolutional-neural-networks","title":"Denovo-GCN: De Novo Peptide Sequencing by Graph Convolutional Neural Networks","methods":"Denovo-GCN","year":2023,"read":true,"reports":["Peptide recall","Amino-acid precision","Amino-acid recall","Precision against coverage"],"rule":"0.1 / 0.5 Da mass rule","il":"Not stated","naming":null,"new_metrics":[],"follows":"DeepNovo (\"This is also the evaluation metric adopted by DeepNovo\")"},{"id":49,"slug":"de-novo-mass-spectrometry-peptide-sequencing-with-a-transformer-model-preprint","title":"De novo mass spectrometry peptide sequencing with a transformer model","methods":"Casanovo","year":2022,"read":true,"reports":["Peptide precision","Amino-acid precision","Precision against coverage","Area under that curve"],"rule":"0.1 / 0.5 Da mass rule","il":"Not stated","naming":null,"new_metrics":[],"follows":"DeepNovo (Tran et al., 2017) definitions"},{"id":50,"slug":"novorank-refinement-for-de-novo-peptide-sequencing-based-on-spectral-clustering","title":"NovoRank: Refinement for De Novo Peptide Sequencing Based on Spectral Clustering and Deep Learning","methods":"NovoRank","year":2025,"read":true,"reports":["Peptide recall","Peptide precision"],"rule":"Exact sequence","il":"Equal","naming":null,"new_metrics":[],"follows":null},{"id":52,"slug":"deps-an-improved-deep-learning-model-for-de-novo-peptide-sequencing","title":"DePS: An improved deep learning model for de novo peptide sequencing","methods":"DePS","year":2022,"read":true,"reports":["Peptide recall","Amino-acid precision","Amino-acid recall"],"rule":"Not stated","il":"Not stated","naming":"recall over predictions","new_metrics":[],"follows":"DeepNovoV2 (\"the same evaluation metrics as those proposed by DeepNovoV2\")"},{"id":53,"slug":"deep-novo-a-plus-improving-the-deep-learning-model-for-de-novo-peptide","title":"Deep Novo A+: Improving the Deep Learning Model for De Novo Peptide Sequencing with Additional Ion Types and Validation Set","methods":"Deep Novo A+","year":2022,"read":true,"reports":[],"rule":"Not stated","il":"Not stated","naming":null,"new_metrics":[],"follows":null},{"id":54,"slug":"computationally-instrument-resolution-independent-de-novo-peptide-sequencing","title":"Computationally instrument-resolution-independent de novo peptide sequencing for high-resolution devices","methods":"PointNovo","year":2021,"read":true,"reports":["Peptide recall","Amino-acid precision","Amino-acid recall"],"rule":"0.1 / 0.5 Da mass rule","il":"Not stated","naming":null,"new_metrics":[],"follows":"DeepNovo and Novor metric (stated)"},{"id":57,"slug":"peptide-sequencing-with-deep-learning","title":"Peptide Sequencing with Deep Learning","methods":"PointNovo","year":2020,"read":true,"reports":["Peptide recall","Amino-acid precision","Amino-acid recall","FDR"],"rule":"0.1 / 0.5 Da mass rule","il":"Not stated","naming":null,"new_metrics":[],"follows":"DeepNovo and Novor evaluation metric (stated explicitly for the PointNovo chapter)"},{"id":59,"slug":"deepnovov2-better-de-novo-peptide-sequencing-with-deep-learning","title":"DeepNovoV2: Better de novo peptide sequencing with deep learning","methods":"DeepNovo V2","year":2019,"read":true,"reports":["Peptide recall","Amino-acid precision","Amino-acid recall"],"rule":"0.1 / 0.5 Da mass rule","il":"Not stated","naming":null,"new_metrics":[],"follows":"DeepNovo (\"we use the same metrics as proposed by DeepNovo\")"},{"id":60,"slug":"deep-learning-enables-de-novo-peptide-sequencing-from-data-independent","title":"Deep learning enables de novo peptide sequencing from data-independent-acquisition mass spectrometry","methods":"DeepNovo-DIA","year":2018,"read":true,"reports":["Peptide recall","Amino-acid precision","FDR","Calibration"],"rule":"Not stated","il":"Not stated","naming":null,"new_metrics":[],"follows":"DeepNovo mass-fragment alignment (ref 11); own script deepnovo_dia_script_test.py"},{"id":62,"slug":"de-novo-peptide-sequencing-by-deep-learning","title":"De novo peptide sequencing by deep learning","methods":"DeepNovo","year":2017,"read":true,"reports":["Peptide recall","Amino-acid precision","Amino-acid recall","Area under that curve"],"rule":"0.1 / 0.5 Da mass rule","il":"Not stated","naming":null,"new_metrics":[],"follows":null},{"id":64,"slug":"causalnovo-advancing-de-novo-peptide-sequencing-via-a-causality-informed","title":"CausalNovo: Advancing De Novo Peptide Sequencing via a Causality-Informed Framework","methods":"CausalNovo","year":2026,"read":true,"reports":["Peptide recall","Amino-acid precision","Amino-acid recall","Area under that curve","PTM precision or recall"],"rule":"Not stated","il":"Not stated","naming":null,"new_metrics":[],"follows":"NovoBench (benchmark setting followed; baseline numbers taken from NovoBench), implied, not stated for the metric code"},{"id":74,"slug":"de-novo-peptide-sequencing-rescoring-and-fdr-estimation-with-winnow","title":"De novo peptide sequencing rescoring and FDR estimation with Winnow","methods":"Winnow","year":2025,"read":true,"reports":["FDR","Calibration"],"rule":"Not stated","il":"Not stated","naming":null,"new_metrics":["Non-parametric (decoy-free) FDR from calibrated confidence"],"follows":"database-grounding introduced in InstaNovo (ref 26)"},{"id":77,"slug":"uncovering-thousands-of-new-peptides-with-sequence-mask-search-hybrid-de-novo","title":"Uncovering Thousands of New Peptides with Sequence-Mask-Search Hybrid De Novo Peptide Sequencing Framework","methods":"SMSNet","year":2019,"read":true,"reports":["Peptide recall","Amino-acid precision","Amino-acid recall"],"rule":"Stricter mass rule","il":"Not stated","naming":null,"new_metrics":[],"follows":null},{"id":93,"slug":"accounting-for-digestion-enzyme-bias-in-casanovo","title":"Accounting for Digestion Enzyme Bias in Casanovo","methods":"Casanovo","year":2024,"read":true,"reports":["Precision against coverage","Area under that curve"],"rule":"Not stated","il":"Not stated","naming":null,"new_metrics":[],"follows":null},{"id":99,"slug":"dpst-de-novo-peptide-sequencing-with-amino-acid-aware-transformers","title":"DPST: De Novo Peptide Sequencing with Amino-Acid-Aware Transformers","methods":"DPST","year":2022,"read":true,"reports":["Peptide recall","Amino-acid precision"],"rule":"Position by position","il":"Not stated","naming":null,"new_metrics":[],"follows":null},{"id":102,"slug":"de-novo-mass-spectrometry-peptide-sequencing-with-a-transformer-model","title":"De novo mass spectrometry peptide sequencing with a transformer model","methods":"Casanovo","year":2022,"read":true,"reports":["Peptide precision","Amino-acid precision","Precision against coverage","Area under that curve"],"rule":"0.1 / 0.5 Da mass rule","il":"Not stated","naming":null,"new_metrics":[],"follows":"Tran et al. 2017 (DeepNovo) nine-species evaluation framework (\"Following (Tran et al., 2017)\")"},{"id":103,"slug":"a-transformer-model-for-de-novo-sequencing-of-data-independent-preprint","title":"A transformer model for de novo sequencing of data independent acquisition mass spectrometry data","methods":"Cascadia","year":2024,"read":true,"reports":["Precision against coverage","Calibration"],"rule":"Not stated","il":"Not stated","naming":null,"new_metrics":["Proteome-match precision-coverage (distinct peptides matching reference proteome, length >= 8)"],"follows":"Tran et al. (DeepNovo-DIA) precursor-level evaluation, as a secondary comparison only"},{"id":104,"slug":"xa-novo-an-accurate-and-high-throughput-mass-spectrometry-based-de-novo-preprint","title":"XA-Novo: an accurate and high-throughput mass spectrometry-based de novo sequencing technology for monoclonal antibodies and antibody mixtures","methods":"XA-Novo","year":2025,"read":true,"reports":["Peptide recall","Amino-acid precision","Amino-acid recall"],"rule":"Not stated","il":"Distinct","naming":null,"new_metrics":[],"follows":null},{"id":105,"slug":"sequence-to-sequence-translation-from-mass-spectra-to-peptides-with-a-preprint","title":"Sequence-to-sequence translation from mass spectra to peptides with a transformer model","methods":"Casanovo","year":2023,"read":true,"reports":["Peptide precision","Amino-acid precision","Precision against coverage","Area under that curve","FDR"],"rule":"0.1 / 0.5 Da mass rule","il":"Not stated","naming":null,"new_metrics":["Length-stratified target-decoy random-matching rate for de novo peptides"],"follows":"Tran et al. (DeepNovo) evaluation (\"Following Tran et al.\")"},{"id":107,"slug":"pi-primenovo-an-accurate-and-efficient-non-autoregressive-deep-preprint","title":"π-PrimeNovo: an accurate and efficient non-autoregressive deep learning model for de novo peptide sequencing","methods":"π-PrimeNovo","year":2024,"read":true,"reports":["Peptide recall","Amino-acid precision","Amino-acid recall","Precision against coverage","FDR","PTM precision or recall"],"rule":"Not stated","il":"Not stated","naming":null,"new_metrics":[],"follows":null},{"id":110,"slug":"pepnet-a-fully-convolutional-neural-network-for-de-novo-peptide-sequencing","title":"PepNet: A Fully Convolutional Neural Network for De novo Peptide Sequencing","methods":"PepNet","year":2022,"read":true,"reports":["Peptide recall","Peptide precision","Amino-acid recall","Precision against coverage"],"rule":"Position by position","il":"Equal","naming":null,"new_metrics":[],"follows":null},{"id":111,"slug":"deep-learning-driven-fragment-ion-series-classification-enables-highly-preprint","title":"Deep learning-driven fragment ion series classification enables highly precise and sensitive de novo peptide sequencing","methods":"Spectralis","year":2023,"read":true,"reports":["Precision against coverage","Sequence similarity"],"rule":"Not stated","il":"Equal","naming":null,"new_metrics":["change precision / change recall"],"follows":null},{"id":114,"slug":"introducing-pandanovo-for-practical-large-scale-de-novo-peptide-sequencing","title":"Introducing PandaNovo for practical large-scale de novo peptide sequencing","methods":"π-HelixNovo","year":2023,"read":true,"reports":["Peptide recall","Peptide precision","Amino-acid precision","Amino-acid recall"],"rule":"0.1 / 0.5 Da mass rule","il":"Not stated","naming":null,"new_metrics":[],"follows":null},{"id":116,"slug":"diffusion-decoding-for-peptide-de-novo-sequencing","title":"Diffusion Decoding for Peptide De Novo Sequencing","methods":"Casanovo-DM2","year":2025,"read":true,"reports":["Peptide precision","Amino-acid precision","Amino-acid recall"],"rule":"Exact sequence","il":"Not stated","naming":null,"new_metrics":[],"follows":"cites Tran 2017, Yilmaz 2022, Mao 2023, Eloff 2025 for the four metrics; default Casanovo hyperparameters"},{"id":119,"slug":"de-novo-peptide-sequencing-for-data-independent-acquisition-dia-using-deep","title":"De Novo Peptide Sequencing for Data-independent Acquisition (DIA) Using Deep Learning","methods":"DiffNovo-DIA","year":2025,"read":false,"reports":[],"rule":null,"il":null,"naming":null,"new_metrics":[],"follows":null},{"id":120,"slug":"transformer-based-de-novo-peptide-sequencing-for-data-independent-acquisition","title":"Transformer-Based De Novo Peptide Sequencing for Data-Independent Acquisition Mass Spectrometry","methods":"DiaTrans","year":2023,"read":true,"reports":["Peptide recall","Amino-acid precision","Amino-acid recall","Precision against coverage","Area under that curve"],"rule":"0.1 / 0.5 Da mass rule","il":"Not stated","naming":"precision over all spectra","new_metrics":[],"follows":"metrics 'consistent with those utilized in DeepNovo-DIA and Casanovo'"},{"id":121,"slug":"deep-active-learning-for-de-novo-peptide-sequencing-from-data-independent","title":"Deep Active Learning for De Novo Peptide Sequencing from Data-independent-acquisition Mass Spectrometry","methods":"AL-DeepNovo-DIA","year":2022,"read":true,"reports":["Peptide recall","Amino-acid recall"],"rule":"Not stated","il":"Not stated","naming":null,"new_metrics":[],"follows":"compares against DeepNovo-DIA's random-selection baseline; no explicit code reuse stated"},{"id":122,"slug":"zero-shot-de-novo-peptide-sequencing-with-open-posttranslational-modification","title":"Zero-shot de novo peptide sequencing with open posttranslational modification discovery","methods":"RNovA","year":2026,"read":true,"reports":["Peptide recall","Amino-acid precision","Amino-acid recall","FDR","PTM precision or recall"],"rule":"0.1 / 0.5 Da mass rule","il":"Not stated","naming":null,"new_metrics":[],"follows":"GraphNovo evaluation metrics ('we adopted the evaluation metrics originally defined in the GraphNovo study')"},{"id":125,"slug":"powernovo2-a-generative-flow-based-approach-to-non-autoregressive-de-novo","title":"PowerNovo2: A generative flow-based approach to non-autoregressive de novo peptide sequencing","methods":"PowerNovo2","year":2026,"read":true,"reports":["Peptide recall","Amino-acid precision","Precision against coverage","Calibration","Sequence similarity"],"rule":"0.1 / 0.5 Da mass rule","il":"Not stated","naming":null,"new_metrics":[],"follows":"Tran et al. (DeepNovo) methodology"},{"id":126,"slug":"punifind-a-unified-large-pre-trained-deep-learning-model-pushing-the-limit-of","title":"pUniFind: a unified large pre-trained deep learning model pushing the limit of mass spectra interpretation","methods":"pUniFind","year":2025,"read":true,"reports":["Peptide recall","FDR","PTM precision or recall"],"rule":"Exact sequence","il":"Not stated","naming":null,"new_metrics":[],"follows":null},{"id":127,"slug":"a-large-scale-unified-deep-learning-model-for-peptide-mass-spectrum","title":"A large-scale unified deep learning model for peptide mass spectrum interpretation trained on multimodal data","methods":"pUniFind","year":2026,"read":true,"reports":["Peptide recall","FDR","PTM precision or recall"],"rule":"Exact sequence","il":"Not stated","naming":null,"new_metrics":[],"follows":null},{"id":161,"slug":"peptide-sequence-determination-from-high-energy-collision-induced-dissociation","title":"Peptide sequence determination from high-energy collision-induced dissociation spectra using artificial neural networks","methods":"Neural-network CID peptide sequencing","year":1995,"read":true,"reports":[],"rule":"Not stated","il":"Distinct","naming":null,"new_metrics":[],"follows":null},{"id":193,"slug":"multienzyme-deep-learning-models-improve-peptide-de-novo-sequencing-by-mass","title":"Multienzyme deep learning models improve peptide de novo sequencing by mass spectrometry proteomics","methods":"Multienzyme DeepNovo models","year":2023,"read":true,"reports":["Peptide recall"],"rule":"Not stated","il":"Not stated","naming":null,"new_metrics":[],"follows":"DeepNovoV2 (metric cited to Tran 2017 and Qiao 2021)"},{"id":196,"slug":"quasinovo-algorithms-for-de-novo-peptide-sequencing","title":"QuasiNovo: Algorithms for De Novo Peptide Sequencing","methods":"QuasiNovo","year":2013,"read":true,"reports":["Amino-acid recall"],"rule":"Not stated","il":"Not stated","naming":null,"new_metrics":["LCSIP prediction accuracy"],"follows":null},{"id":202,"slug":"memnovo-look-back-at-the-spectrum-for-balanced-de-novo-peptide-sequencing-from","title":"MemNovo: Look Back at the Spectrum for Balanced De Novo Peptide Sequencing from Mass Spectrometry","methods":"MemNovo","year":2026,"read":true,"reports":["Peptide recall","Amino-acid precision","Amino-acid recall"],"rule":"Exact sequence","il":"Equal","naming":null,"new_metrics":[],"follows":null},{"id":218,"slug":"foundation-model-for-mass-spectrometry-proteomics","title":"Foundation model for mass spectrometry proteomics","methods":"Casanovo Foundation","year":2025,"read":true,"reports":[],"rule":"Not stated","il":"Not stated","naming":null,"new_metrics":[],"follows":null},{"id":219,"slug":"pi-helixnovo2-making-accurate-online-de-novo-peptide-sequencing-available-to-all","title":"π-HelixNovo2: Making Accurate Online De Novo Peptide Sequencing Available to All","methods":"π-HelixNovo2","year":2026,"read":true,"reports":["Peptide precision","Amino-acid precision","Amino-acid recall"],"rule":"0.1 / 0.5 Da mass rule","il":"Not stated","naming":"recall over predictions","new_metrics":[],"follows":null},{"id":221,"slug":"false-discovery-rate-control-for-trustworthy-ai-based-de-novo-peptide-sequencing","title":"False discovery rate control for trustworthy AI-based de novo peptide sequencing","methods":"FDR control for AI-based de novo sequencing","year":2026,"read":true,"reports":["FDR"],"rule":"Not stated","il":"Not stated","naming":null,"new_metrics":["Counterpart Calibration Theory (CCT) FDR estimate","isotopologue-missingness empirical FDR"],"follows":null},{"id":222,"slug":"regressor-guided-diffusion-model-for-de-novo-peptide-sequencing-with-preprint","title":"Regressor-guided Diffusion Model for De Novo Peptide Sequencing with Explicit Mass Control","methods":"DiffuNovo","year":2026,"read":true,"reports":["Peptide recall","Amino-acid precision","Area under that curve","PTM precision or recall","Mass-based"],"rule":"Not stated","il":"Not stated","naming":null,"new_metrics":["Plausible prediction rate","Mass error (MAE)"],"follows":null},{"id":225,"slug":"protein-identification-with-deep-learning-from-abc-to-xyz","title":"Protein identification with deep learning: from abc to xyz","methods":"Deep-learning protein identification primer","year":2017,"read":true,"reports":["Peptide recall","Amino-acid recall","FDR"],"rule":"Not stated","il":"Not stated","naming":null,"new_metrics":[],"follows":"DeepNovo"},{"id":251,"slug":"instanovo-p-a-de-novo-peptide-sequencing-model-for-phosphoproteomics","title":"InstaNovo-P: a de novo peptide sequencing model for phosphoproteomics","methods":"InstaNovo-P","year":2026,"read":true,"reports":["Peptide recall","Peptide precision","Area under that curve","FDR","PTM precision or recall"],"rule":"Exact sequence","il":"Equal","naming":null,"new_metrics":[],"follows":"InstaNovo"},{"id":254,"slug":"discrete-diffusion-with-physical-mass-constraints-for-de-novo-peptide-sequencing","title":"Discrete Diffusion with Physical Mass Constraints for De Novo Peptide Sequencing","methods":"PhysNovo","year":2026,"read":true,"reports":["Peptide recall","Amino-acid precision","Precision against coverage","Area under that curve","PTM precision or recall"],"rule":"0.1 / 0.5 Da mass rule","il":"Equal","naming":"precision over all spectra","new_metrics":[],"follows":"NovoBench protocol ('Consistent with standard protocols (Zhou et al., 2024)'; datasets Nine-species, Seven-species, HC-PT)"},{"id":255,"slug":"prime-diffnovo-accurate-peptide-de-novo-sequencing-via-non-autoregressive","title":"Prime-DiffNovo: Accurate Peptide De Novo Sequencing via Non-autoregressive Generation and Diffusion Refinement","methods":"Prime-DiffNovo","year":2026,"read":true,"reports":["Peptide recall","Amino-acid precision","Amino-acid recall"],"rule":"0.1 / 0.5 Da mass rule","il":"Equal","naming":null,"new_metrics":[],"follows":"Casanovo (AA recall computation)"},{"id":264,"slug":"accurate-and-ultra-fast-de-novo-hla-i-immunopeptide-sequencing-with-foxnovo","title":"Accurate and ultra-fast de novo HLA-I immunopeptide sequencing with FoxNovo","methods":"FoxNovo","year":2026,"read":true,"reports":["Peptide recall","Precision against coverage","Sequence similarity"],"rule":"Exact sequence","il":"Equal","naming":null,"new_metrics":[],"follows":null},{"id":265,"slug":"tandem-mass-spectra-representation-design-for-transformer-based-de-novo-peptide","title":"Tandem Mass Spectra Representation Design for Transformer-Based De Novo Peptide Sequencing","methods":"Spectrum representation ablation study","year":2026,"read":true,"reports":["Peptide recall","Amino-acid precision","Amino-acid recall"],"rule":"Not stated","il":"Equal","naming":null,"new_metrics":[],"follows":"Casanovo evaluation code and thresholds"},{"id":269,"slug":"pairwise-attention-leveraging-mass-differences-to-enhance-de-novo-sequencing-of","title":"Pairwise Attention: Leveraging Mass Differences to Enhance De Novo Sequencing of Mass Spectra","methods":"Pairwise","year":2025,"read":true,"reports":["Peptide recall","Precision against coverage"],"rule":"0.1 / 0.5 Da mass rule","il":"Not stated","naming":null,"new_metrics":[],"follows":"DeepNovo amino-acid matching; Casanovo peptide score"},{"id":270,"slug":"modanovo-a-unified-model-for-post-translational-modification-aware-de-novo","title":"Modanovo: A Unified Model for Post-translational Modification-Aware De Novo Sequencing Using Experimental Spectra From In Vivo and Synthetic Peptides","methods":"Modanovo","year":2025,"read":true,"reports":["Peptide recall","Precision against coverage","Area under that curve"],"rule":"Exact sequence","il":"Equal","naming":null,"new_metrics":[],"follows":"Casanovo split strategy and score; Spectralis-style evaluation (same lab)"},{"id":271,"slug":"improvements-to-casanovo-a-deep-learning-de-novo-peptide-sequencer","title":"Improvements to Casanovo, a Deep Learning De Novo Peptide Sequencer","methods":"Casanovo","year":2025,"read":true,"reports":["Area under that curve","FDR","Calibration"],"rule":"Not stated","il":"Equal","naming":null,"new_metrics":[],"follows":"Casanovo (own evaluate option)"},{"id":273,"slug":"learning-from-tandem-mass-spectra-at-scale-with-a-self-supervised-foundation","title":"Learning from tandem mass spectra at scale with a self-supervised foundation model for proteomics","methods":"InstaNovo-FM","year":2026,"read":true,"reports":["Peptide recall"],"rule":"Not stated","il":"Not stated","naming":null,"new_metrics":[],"follows":"InstaNovo benchmark datasets"},{"id":283,"slug":"biatnovo-a-self-attention-based-bidirectional-peptide-sequencing-method","title":"BiATNovo: A Self-Attention based Bidirectional Peptide Sequencing Method","methods":"BiATNovo","year":2023,"read":true,"reports":["Peptide recall","Amino-acid precision","Sequence similarity"],"rule":"Not stated","il":"Not stated","naming":null,"new_metrics":["Position-BLEU","Alignment score"],"follows":"DeepNovo-DIA (baseline and test sets)"},{"id":302,"slug":"dinovo-enables-high-coverage-and-high-confidence-de-novo-peptide-sequencing-via","title":"DiNovo enables high-coverage and high-confidence de novo peptide sequencing via mirror proteases and deep learning","methods":"DiNovo","year":2026,"read":true,"reports":["FDR"],"rule":"Not stated","il":"Not stated","naming":null,"new_metrics":["TD-mapping FDR for de novo peptides"],"follows":null},{"id":352,"slug":"universal-biological-sequence-reranking-for-improved-de-novo-peptide-preprint","title":"Universal Biological Sequence Reranking for Improved De Novo Peptide Sequencing","methods":"ByNovo, RankNovo","year":2025,"read":true,"reports":["Peptide recall","Amino-acid precision"],"rule":"0.1 / 0.5 Da mass rule","il":"Not stated","naming":null,"new_metrics":["PMD (Peptide Mass Deviation)","RMD (Residual Mass Deviation)"],"follows":"DeepNovo/Casanovo-style residue matching (not attributed explicitly)"},{"id":353,"slug":"pi-mnovo-improves-de-novo-peptide-sequencing-through-microbial-domain","title":"π-MNovo improves de novo peptide sequencing through microbial-domain adaptation and evidence-guided candidate selection","methods":"π-MNovo","year":2026,"read":true,"reports":["Peptide recall","Precision against coverage","Calibration"],"rule":"0.1 / 0.5 Da mass rule","il":"Equal","naming":null,"new_metrics":[],"follows":null},{"id":354,"slug":"gyronovo-error-guided-fragment-imputation-with-mass-aware-attention-for-de-novo","title":"GyroNovo: Error-Guided Fragment Imputation with Mass-Aware Attention for De Novo Peptide Sequencing","methods":"GyroNovo","year":2026,"read":true,"reports":["Peptide recall","Amino-acid precision","Area under that curve","PTM precision or recall"],"rule":"0.1 / 0.5 Da mass rule","il":"Not stated","naming":null,"new_metrics":[],"follows":"NovoBench / LIPNovo (reproduced LIPNovo baseline); Tran et al. 2017 residue matching"},{"id":359,"slug":"novorank-machine-learning-based-post-processing-for-performance-improvement-in","title":"NovoRank: Machine Learning Based Post-processing for Performance Improvement in De Novo Peptide Sequencing","methods":"NovoRank","year":2022,"read":true,"reports":["Peptide recall","Amino-acid recall"],"rule":"Not stated","il":"Not stated","naming":null,"new_metrics":[],"follows":null},{"id":432,"slug":"lipnovo-plus-self-reflective-latent-imputation-for-robust-de-novo-peptide","title":"LIPNovo+: Self-Reflective Latent Imputation for Robust De Novo Peptide Sequencing","methods":"LIPNovo, LIPNovo+","year":2026,"read":true,"reports":["Peptide recall","Amino-acid precision","Area under that curve","PTM precision or recall"],"rule":"Not stated","il":"Not stated","naming":null,"new_metrics":[],"follows":"NovoBench"},{"id":433,"slug":"rt-gctnovo-a-peptide-de-novo-sequencing-model-incorporating-gated-multi-scale","title":"RT-GCTnovo: A Peptide De Novo Sequencing Model Incorporating Gated Multi-scale Features and Dynamic Mass Masks","methods":"RT-GCTnovo","year":2025,"read":true,"reports":["Peptide precision","Amino-acid precision"],"rule":"0.1 / 0.5 Da mass rule","il":"Not stated","naming":null,"new_metrics":[],"follows":"DeepNovo multispecies (nine-species) benchmark"},{"id":434,"slug":"protein-language-model-aligned-spectra-embeddings-for-de-novo-peptide-sequencing","title":"Protein Language Model-Aligned Spectra Embeddings for De Novo Peptide Sequencing","methods":"PLMNovo","year":2025,"read":true,"reports":["Peptide recall","Amino-acid precision"],"rule":"Not stated","il":"Not stated","naming":null,"new_metrics":[],"follows":null},{"id":438,"slug":"curriculum-learning-for-biological-sequence-prediction-the-case-of-de-novo","title":"Curriculum Learning for Biological Sequence Prediction: The Case of De Novo Peptide Sequencing","methods":"RefineNovo","year":2025,"read":true,"reports":["Peptide recall","Amino-acid precision"],"rule":"Not stated","il":"Not stated","naming":null,"new_metrics":[],"follows":"Casanovo (Yilmaz et al. 2022) matching criteria"},{"id":439,"slug":"latent-imputation-before-prediction-a-new-computational-paradigm-for-de-novo","title":"Latent Imputation before Prediction: A New Computational Paradigm for De Novo Peptide Sequencing","methods":"LIPNovo","year":2025,"read":true,"reports":["Peptide recall","Amino-acid precision","Area under that curve","PTM precision or recall"],"rule":"Not stated","il":"Not stated","naming":null,"new_metrics":[],"follows":"NovoBench (baseline numbers 'provided by NovoBench'; Casanovo retrained)"},{"id":440,"slug":"adanovo-towards-robust-de-novo-peptide-sequencing-in-proteomics-against-data","title":"AdaNovo: Towards Robust De Novo Peptide Sequencing in Proteomics against Data Biases","methods":"AdaNovo","year":2024,"read":true,"reports":["Peptide recall","Amino-acid precision","PTM precision or recall"],"rule":"0.1 / 0.5 Da mass rule","il":"Not stated","naming":"precision over all spectra","new_metrics":[],"follows":"DeepNovo / Casanovo methodology"},{"id":482,"slug":"a-residual-network-for-de-novo-peptide-sequencing-with-attention-mechanism","title":"A residual network for de novo peptide sequencing with attention mechanism","methods":"RANovo","year":2020,"read":false,"reports":[],"rule":null,"il":null,"naming":null,"new_metrics":[],"follows":null},{"id":494,"slug":"uncovering-thousands-of-new-hla-antigens-and-phosphopeptides-with-deep-learning","title":"Uncovering thousands of new HLA antigens and phosphopeptides with deep learning-based sequence-mask-search de novo peptide sequencing framework","methods":"SMSNet","year":2019,"read":true,"reports":["Peptide recall","Amino-acid recall","Precision against coverage"],"rule":"Stricter mass rule","il":"Not stated","naming":null,"new_metrics":[],"follows":null},{"id":545,"slug":"dyconovo-a-de-novo-peptide-prediction-model-based-on-dynamic-convolution-and","title":"DyCoNovo: a De Novo Peptide Prediction Model Based on Dynamic Convolution and Phased Contrastive Learning","methods":"DyCoNovo","year":2025,"read":false,"reports":[],"rule":null,"il":null,"naming":null,"new_metrics":[],"follows":null},{"id":551,"slug":"critical-evaluation-of-the-use-of-artificial-data-for-machine-learning-based-de","title":"Critical evaluation of the use of artificial data for machine learning based de novo peptide identification","methods":"Artificial training data evaluation for ML de novo","year":2023,"read":true,"reports":["Peptide recall","Amino-acid precision"],"rule":"0.1 / 0.5 Da mass rule","il":"Not stated","naming":null,"new_metrics":[],"follows":"PointNovo"},{"id":641,"slug":"generative-ai-models-for-the-protein-scaffold-filling-problem","title":"Generative AI Models for the Protein Scaffold Filling Problem","methods":"Generative AI protein scaffold filling","year":2025,"read":false,"reports":[],"rule":null,"il":null,"naming":null,"new_metrics":[],"follows":null},{"id":651,"slug":"deep-learning-approaches-for-the-protein-scaffold-filling-problem","title":"Deep Learning Approaches for the Protein Scaffold Filling Problem","methods":"Deep learning protein scaffold filling","year":2022,"read":false,"reports":[],"rule":null,"il":null,"naming":null,"new_metrics":[],"follows":null},{"id":695,"slug":"tf-gcnnovo-a-peptide-sequence-prediction-model-integrating-transformer-and","title":"TF-GCNNovo: A Peptide Sequence Prediction Model Integrating Transformer and Graph Convolutional Network","methods":"TF-GCNNovo","year":2025,"read":false,"reports":[],"rule":null,"il":null,"naming":null,"new_metrics":[],"follows":null},{"id":752,"slug":"bert-model-for-de-novo-protein-sequencing-using-tandem-mass-spectra","title":"BERT model for de novo protein sequencing using tandem mass spectra","methods":"PowerNovo","year":2024,"read":true,"reports":["Peptide recall"],"rule":"Not stated","il":"Not stated","naming":null,"new_metrics":[],"follows":null},{"id":753,"slug":"evaluation-of-the-accuracy-of-false-alarm-frequency-control-methods-for-de-novo","title":"Evaluation of the accuracy of false alarm frequency control methods for de novo spectrum","methods":"PepNet vs Tide false-discovery evaluation","year":2025,"read":true,"reports":["Mass-based","Sequence similarity"],"rule":"Not stated","il":"Not stated","naming":null,"new_metrics":[],"follows":null},{"id":978,"slug":"universal-biological-sequence-reranking-for-improved-de-novo-peptide-sequencing","title":"Universal Biological Sequence Reranking for Improved De Novo Peptide Sequencing","methods":"ByNovo, RankNovo","year":2025,"read":true,"reports":["Peptide recall","Amino-acid precision"],"rule":"0.1 / 0.5 Da mass rule","il":"Not stated","naming":null,"new_metrics":["PMD (Peptide Mass Deviation)","RMD (Residual Mass Deviation)"],"follows":"DeepNovo/Casanovo-style residue matching (not attributed explicitly)"},{"id":979,"slug":"bidirectional-representations-augmented-autoregressive-biological-sequence","title":"Bidirectional Representations Augmented Autoregressive Biological Sequence Generation","methods":"CrossNovo","year":2025,"read":true,"reports":["Peptide recall","Amino-acid precision","Precision against coverage"],"rule":"0.1 / 0.5 Da mass rule","il":"Not stated","naming":"precision over all spectra","new_metrics":[],"follows":"'Following standard practice' (Casanovo/PrimeNovo protocol; MassIVE-KB training)"},{"id":981,"slug":"dion-fragmentation-based-invariance-for-self-supervised-learning-of-tandem-mass","title":"dIon: Fragmentation-Based Invariance for Self-Supervised Learning of Tandem Mass Spectra","methods":"dIon","year":2026,"read":true,"reports":["Peptide recall","Amino-acid precision","Precision against coverage"],"rule":"0.1 / 0.5 Da mass rule","il":"Equal","naming":null,"new_metrics":[],"follows":"Casanovo implementation of the DeepNovo matcher, applied uniformly to all models (incl. InstaNovo 1.2.2 and Casanovo 5.2.1)"},{"id":984,"slug":"pretrained-spectrum-representations-for-diverse-tandem-mass-spectrometry","title":"Pretrained spectrum representations for diverse tandem mass spectrometry proteomics tasks","methods":"Casanovo Foundation","year":2026,"read":true,"reports":[],"rule":"Not stated","il":"Not stated","naming":null,"new_metrics":[],"follows":null},{"id":989,"slug":"massnet-billion-scale-ai-ready-mass-spectral-corpus-empowering-scalable-deep","title":"MassNet: billion-scale AI-ready mass spectral corpus empowering scalable deep learning in proteomics","methods":"XuanjiNovo","year":2026,"read":true,"reports":["Peptide recall","Amino-acid recall","Area under that curve","FDR"],"rule":"Not stated","il":"Not stated","naming":null,"new_metrics":[],"follows":null},{"id":990,"slug":"scopenovo-a-de-novo-peptide-sequencing-method-using-spectrum-spectrum","title":"ScopeNovo: a De Novo Peptide Sequencing Method Using Spectrum-Spectrum Similarity to Refine the Prediction Accuracy","methods":"ScopeNovo","year":2026,"read":true,"reports":["Peptide recall","Precision against coverage"],"rule":"Not stated","il":"Not stated","naming":null,"new_metrics":[],"follows":null},{"id":996,"slug":"peppr-de-novo-peptide-sequencing-with-peptide-priors","title":"PepPr: De Novo peptide sequencing with peptide priors","methods":"PepPr","year":2026,"read":true,"reports":["Peptide recall","Area under that curve"],"rule":"0.1 / 0.5 Da mass rule","il":"Equal","naming":null,"new_metrics":[],"follows":"Casanovo definitions (coverage against all spectra)"},{"id":998,"slug":"novocrf-globally-normalized-de-novo-peptide-sequencing-on-a-prefix-mass-lattice","title":"NovoCRF: Globally Normalized De Novo Peptide Sequencing on a Prefix-Mass Lattice","methods":"NovoCRF","year":2026,"read":true,"reports":["Peptide recall","Amino-acid recall","Area under that curve","Calibration"],"rule":"0.1 / 0.5 Da mass rule","il":"Equal","naming":null,"new_metrics":["recall by fragment evidence"],"follows":"Casanovo-style matching; denovo_benchmarks container for baselines"},{"id":1000,"slug":"pi-specformer-a-mass-spectrometry-based-foundation-model-for-decoding-the-dark","title":"π-SPECFormer: A Mass Spectrometry-based Foundation Model for Decoding the Dark Proteome","methods":"π-SPECFormer","year":2026,"read":true,"reports":["Peptide recall","Amino-acid precision"],"rule":"Exact sequence","il":"Not stated","naming":null,"new_metrics":[],"follows":null},{"id":1003,"slug":"casanovo-tmt-deep-learning-based-de-novo-sequencing-of-tmtpro-labeled-peptides","title":"Casanovo-TMT: deep learning-based de novo sequencing of TMTpro-labeled peptides and phosphopeptides","methods":"Casanovo-TMT","year":2026,"read":true,"reports":["Peptide recall","FDR"],"rule":"Exact sequence","il":"Equal","naming":"precision over all spectra","new_metrics":["decoy-digest threshold for de novo output"],"follows":"Casanovo 5.2.0 (fine-tuned), with its own stricter evaluator"},{"id":1004,"slug":"seminovo-learning-beyond-search-identified-spectra-for-de-novo-peptide","title":"SemiNovo: Learning Beyond Search-Identified Spectra for De Novo Peptide Sequencing","methods":"SemiNovo","year":2026,"read":true,"reports":["Peptide recall","Amino-acid precision","Area under that curve","PTM precision or recall"],"rule":"0.1 / 0.5 Da mass rule","il":"Equal","naming":"precision over all spectra","new_metrics":[],"follows":"NovoBench official implementation"},{"id":1005,"slug":"encnovo-de-novo-peptide-sequencing-with-an-encoder-only-architectures","title":"EncNovo: De Novo Peptide Sequencing With an Encoder-Only Architectures","methods":"EncNovo","year":2026,"read":false,"reports":[],"rule":null,"il":null,"naming":null,"new_metrics":[],"follows":null}],"code":[{"tool":"DeepNovo","versions":"all (since 0.0.1, 2017)","lineage":"origin","rule":"0.1 / 0.5 Da, forward only","order":"correct","empty":"kept","peptide":"correct / all labelled spectra; a peptide counts when every TRUE residue matched, whatever the prediction's length","curve":"none","repo":"nh2tran/DeepNovo","ref":"8b0ec83","path":"deepnovo_worker_test.py","url":"https://github.com/nh2tran/DeepNovo/blob/8b0ec83/deepnovo_worker_test.py"},{"tool":"DeepNovo-DIA, DeepNovoV2, PointNovo, PGPointNovo, DeepNovoAA","versions":"all","lineage":"DeepNovo","rule":"0.1 / 0.5 Da, forward only","order":"correct","empty":"kept","peptide":"as DeepNovo, plus a peptide precision over predicted features; with more than one candidate per feature the evaluator keeps the one closest to the truth","curve":"score cut-off at a target AA accuracy","repo":"irleader/PointNovo","ref":"552b09a","path":"deepnovo_worker_test.py","url":"https://github.com/irleader/PointNovo/blob/552b09a/deepnovo_worker_test.py"},{"tool":"Kaiko","versions":"all","lineage":"DeepNovo","rule":"0.1 / 0.5 Da, forward only; also a position-by-position token accuracy","order":"correct","empty":"kept","peptide":"as DeepNovo","curve":"none","repo":"PNNL-Comp-Mass-Spec/Kaiko","ref":"9aa0c39","path":"src/deepnovo_main_modules.py","url":"https://github.com/PNNL-Comp-Mass-Spec/Kaiko/blob/9aa0c39/src/deepnovo_main_modules.py"},{"tool":"SMSNet","versions":"all","lineage":"independent","rule":"exact residue mass (< 0.0001 Da) and 0.03 Da prefix, forward only","order":"recall only","empty":"kept","peptide":"exact string, L read as I on the label only","curve":"none","repo":"cmb-chula/SMSNet","ref":"facfaf4","path":"nmt/utils/evaluation_utils.py","url":"https://github.com/cmb-chula/SMSNet/blob/facfaf4/nmt/utils/evaluation_utils.py"},{"tool":"PepNet","versions":"all","lineage":"independent","rule":"exact string, I read as L","order":"no AA metric","empty":"kept","peptide":"correct / all rows","curve":"none","repo":"lkytal/PepNet","ref":"1ef0de5","path":"evaluation.py","url":"https://github.com/lkytal/PepNet/blob/1ef0de5/evaluation.py"},{"tool":"Casanovo","versions":"v1.1.0 to v2.1.1","lineage":"reference (copied from an unmerged depthcharge branch)","rule":"0.1 / 0.5 Da, one pass on the mass before or after each residue","order":"correct","empty":"dropped","peptide":"correct / spectra with a well-formed prediction","curve":"none","repo":"Noble-Lab/casanovo","ref":"3fa0556","path":"casanovo/denovo/evaluate.py","url":"https://github.com/Noble-Lab/casanovo/blob/3fa0556/casanovo/denovo/evaluate.py"},{"tool":"Casanovo","versions":"v3.0.0 to v3.5.0","lineage":"reference","rule":"0.1 / 0.5 Da, prefix then suffix","order":"swapped","empty":"dropped","peptide":"correct / spectra with a well-formed prediction","curve":"none","repo":"Noble-Lab/casanovo","ref":"ea2e875","path":"casanovo/denovo/model.py","url":"https://github.com/Noble-Lab/casanovo/blob/ea2e875/casanovo/denovo/model.py"},{"tool":"Casanovo","versions":"v4.0.0 to v4.3.0","lineage":"reference","rule":"0.1 / 0.5 Da, prefix then suffix","order":"correct","empty":"kept, but in validation an empty spectrum shifts the rest of its batch onto the wrong truth","peptide":"correct / all spectra","curve":"none","repo":"Noble-Lab/casanovo","ref":"3c2d3f5","path":"casanovo/denovo/model.py","url":"https://github.com/Noble-Lab/casanovo/blob/3c2d3f5/casanovo/denovo/model.py"},{"tool":"Casanovo","versions":"v5.0.0 to v5.1.2","lineage":"reference","rule":"0.1 / 0.5 Da, prefix then suffix; validation splits M[Oxidation] into two massless pieces","order":"correct","empty":"kept","peptide":"sequence --evaluate prints 0, because padding is counted as residues","curve":"none","repo":"Noble-Lab/casanovo","ref":"c01d818","path":"casanovo/denovo/model_runner.py","url":"https://github.com/Noble-Lab/casanovo/blob/c01d818/casanovo/denovo/model_runner.py"},{"tool":"Casanovo","versions":"v5.2.0 and later","lineage":"reference","rule":"0.1 / 0.5 Da, prefix then suffix; validation still splits bracketed modifications","order":"correct","empty":"kept","peptide":"printed as peptide precision, computed as correct / all spectra","curve":"none released; the dev branch adds precision against coverage","repo":"Noble-Lab/casanovo","ref":"40f4e97","path":"casanovo/denovo/evaluate.py","url":"https://github.com/Noble-Lab/casanovo/blob/40f4e97/casanovo/denovo/evaluate.py"},{"tool":"ContraNovo","versions":"latest","lineage":"Casanovo v3.3","rule":"0.1 / 0.5 Da, prefix then suffix","order":"swapped","empty":"dropped","peptide":"correct / spectra with a well-formed prediction","curve":"none","repo":"BEAM-Labs/ContraNovo","ref":"503fdb3","path":"ContraNovo/denovo/model.py","url":"https://github.com/BEAM-Labs/ContraNovo/blob/503fdb3/ContraNovo/denovo/model.py"},{"tool":"PrimeNovo, π-PrimeNovo, ScopeNovo","versions":"latest (every copy)","lineage":"Casanovo v3.2","rule":"0.1 / 0.5 Da, prefix then suffix","order":"swapped","empty":"dropped","peptide":"correct / spectra with a well-formed prediction","curve":"none","repo":"BEAM-Labs/denovo","ref":"437f3c7","path":"PrimeNovo/denovo/model.py","url":"https://github.com/BEAM-Labs/denovo/blob/437f3c7/PrimeNovo/denovo/model.py"},{"tool":"RefineNovo","versions":"latest","lineage":"Casanovo v3.2","rule":"0.1 / 0.5 Da, prefix then suffix; plus a by-position per-residue score for M(ox), F, Q, K","order":"swapped","empty":"dropped","peptide":"correct / spectra with a well-formed prediction","curve":"none","repo":"BEAM-Labs/denovo","ref":"437f3c7","path":"RefineNovo/RefineNovo/denovo/model.py","url":"https://github.com/BEAM-Labs/denovo/blob/437f3c7/RefineNovo/RefineNovo/denovo/model.py"},{"tool":"RankNovo","versions":"latest","lineage":"Casanovo v3.3","rule":"0.1 / 0.5 Da, prefix then suffix","order":"swapped","empty":"n/a (reranks given candidates)","peptide":"unweighted mean of per-batch values","curve":"none; PMD and RMD are training targets, not reported","repo":"BEAM-Labs/denovo","ref":"437f3c7","path":"RankNovo/RankNovo/denovo/model_runner.py","url":"https://github.com/BEAM-Labs/denovo/blob/437f3c7/RankNovo/RankNovo/denovo/model_runner.py"},{"tool":"XuanjiNovo (MassNet-DDA)","versions":"v1.1","lineage":"Casanovo v3.2","rule":"0.1 / 0.5 Da, prefix then suffix","order":"swapped","empty":"dropped","peptide":"correct / spectra with a well-formed prediction","curve":"none","repo":"guomics-lab/MassNet-DDA","ref":"cf1432a","path":"XuanjiNovo/denovo/model.py","url":"https://github.com/guomics-lab/MassNet-DDA/blob/cf1432a/XuanjiNovo/denovo/model.py"},{"tool":"XuanjiNovo (MassNet-DDA)","versions":"current","lineage":"Casanovo v3.2","rule":"0.1 / 0.5 Da, prefix then suffix","order":"correct","empty":"dropped","peptide":"correct / spectra with a well-formed prediction","curve":"none","repo":"guomics-lab/MassNet-DDA","ref":"12ecd75","path":"XuanjiNovo/denovo/model.py","url":"https://github.com/guomics-lab/MassNet-DDA/blob/12ecd75/XuanjiNovo/denovo/model.py"},{"tool":"DiffNovo","versions":"latest","lineage":"Casanovo v3.2","rule":"0.1 / 0.5 Da, prefix then suffix","order":"swapped","empty":"dropped","peptide":"the better of two decoders per batch, and the peptide slot holds an amino-acid value","curve":"none","repo":"Biocomputing-Research-Group/DiffNovo","ref":"bf9ac9f","path":"diffnovo/denovo/model.py","url":"https://github.com/Biocomputing-Research-Group/DiffNovo/blob/bf9ac9f/diffnovo/denovo/model.py"},{"tool":"Transformer-DIA, DiffNovo-DIA","versions":"latest","lineage":"Casanovo v3.2","rule":"0.1 / 0.5 Da, prefix then suffix","order":"swapped","empty":"dropped","peptide":"correct / spectra with a well-formed prediction","curve":"none","repo":"Biocomputing-Research-Group/Transformer-DIA","ref":"1530fc3","path":"casanovo_dia/denovo/model.py","url":"https://github.com/Biocomputing-Research-Group/Transformer-DIA/blob/1530fc3/casanovo_dia/denovo/model.py"},{"tool":"π-HelixNovo","versions":"every tag","lineage":"Casanovo v3.0","rule":"0.1 / 0.5 Da, prefix then suffix","order":"swapped","empty":"dropped","peptide":"correct / spectra with a well-formed prediction","curve":"none","repo":"PHOENIXcenter/pi-HelixNovo","ref":"49c0673","path":"denovo/model.py","url":"https://github.com/PHOENIXcenter/pi-HelixNovo/blob/49c0673/denovo/model.py"},{"tool":"π-MNovo","versions":"v0.1.0","lineage":"Casanovo v3.2","rule":"0.1 / 0.5 Da, prefix then suffix","order":"swapped","empty":"dropped","peptide":"correct / spectra with a well-formed prediction","curve":"none","repo":"PHOENIXcenter/pi-MNovo","ref":"636ca9c","path":"MNovo/denovo/model.py","url":"https://github.com/PHOENIXcenter/pi-MNovo/blob/636ca9c/MNovo/denovo/model.py"},{"tool":"π-MNovo","versions":"v0.2.0 and later","lineage":"Casanovo v3.2, rewritten","rule":"0.1 / 0.5 Da, prefix then suffix, I read as L","order":"correct","empty":"kept","peptide":"correct / all spectra, pooled over batches","curve":"none","repo":"PHOENIXcenter/pi-MNovo","ref":"8f6ece7","path":"MNovo/denovo/metric_counts.py","url":"https://github.com/PHOENIXcenter/pi-MNovo/blob/8f6ece7/MNovo/denovo/metric_counts.py"},{"tool":"TSARseqNovo","versions":"latest","lineage":"Casanovo v4.0","rule":"0.1 / 0.5 Da, prefix then suffix","order":"swapped","empty":"n/a (always predicts)","peptide":"correct / all spectra","curve":"none","repo":"qiyueliuhuo8/TSARseqNovo","ref":"0bd762c","path":"denovo/model/model_runner.py","url":"https://github.com/qiyueliuhuo8/TSARseqNovo/blob/0bd762c/denovo/model/model_runner.py"},{"tool":"AdaNovo","versions":"latest","lineage":"Casanovo v3.3","rule":"0.1 / 0.5 Da, prefix then suffix; PTM precision and recall","order":"correct","empty":"kept","peptide":"correct / all spectra","curve":"none","repo":"Westlake-OmicsAI/adanovo_v1","ref":"212fbd4","path":"denovo/model.py","url":"https://github.com/Westlake-OmicsAI/adanovo_v1/blob/212fbd4/denovo/model.py"},{"tool":"Modanovo, PLMNovo","versions":"latest","lineage":"Casanovo v4.0","rule":"0.1 / 0.5 Da, prefix then suffix","order":"correct","empty":"as Casanovo v4","peptide":"correct / all spectra","curve":"none","repo":"gagneurlab/Modanovo","ref":"7e02cec","path":"modanovo/denovo/model.py","url":"https://github.com/gagneurlab/Modanovo/blob/7e02cec/modanovo/denovo/model.py"},{"tool":"Pairwise, dIon","versions":"latest","lineage":"Casanovo v4.1","rule":"0.1 / 0.5 Da, prefix then suffix","order":"correct","empty":"kept (dIon); dropped from coverage (Pairwise ROC script)","peptide":"correct / all spectra","curve":"precision against coverage; dIon over all spectra, Pairwise over predicted ones","repo":"statisticalbiotechnology/dIon","ref":"4add96e","path":"scripts/evaluate_released_casanovo_mztab.py","url":"https://github.com/statisticalbiotechnology/dIon/blob/4add96e/scripts/evaluate_released_casanovo_mztab.py"},{"tool":"NovoBench (and LIPNovo, SemiNovo)","versions":"latest","lineage":"Casanovo v3.2, rewritten","rule":"0.1 / 0.5 Da, prefix then suffix; strict tokeniser, no N-terminal modifications","order":"correct","empty":"kept","peptide":"printed as peptide precision, computed as correct / all spectra","curve":"area under precision against RECALL","repo":"jingbo02/NovoBench","ref":"d1eff5b","path":"novobench/metrics/evaluate.py","url":"https://github.com/jingbo02/NovoBench/blob/d1eff5b/novobench/metrics/evaluate.py"},{"tool":"MemNovo","versions":"latest","lineage":"Casanovo v5.0 wrapper","rule":"modifications stripped; glycine missing from the mass table; other model names fall back to unordered letter counts","order":"correct","empty":"dropped from precision","peptide":"correct / all predictions","curve":"none","repo":"AIMS-Lab-HKUSTGZ/MemNovo","ref":"7c7565a","path":"evaluation/evaluator.py","url":"https://github.com/AIMS-Lab-HKUSTGZ/MemNovo/blob/7c7565a/evaluation/evaluator.py"},{"tool":"Cascadia","versions":"latest","lineage":"independent","rule":"exact token, teacher-forced (I and L distinct)","order":"accuracy only","empty":"n/a","peptide":"every position right under teacher forcing","curve":"proteome-match precision against coverage (in the paper)","repo":"Noble-Lab/cascadia","ref":"b55a177","path":"cascadia/model.py","url":"https://github.com/Noble-Lab/cascadia/blob/b55a177/cascadia/model.py"},{"tool":"DpNovo","versions":"latest","lineage":"independent","rule":"character edit distance, I = L","order":"correct","empty":"n/a","peptide":"exact position-wise identity","curve":"none","repo":"4chuanMcChicken/DpNovo","ref":"cf5e1eb","path":"DpNovo.ipynb","url":"https://github.com/4chuanMcChicken/DpNovo/blob/cf5e1eb/DpNovo.ipynb"},{"tool":"InstaNovo","versions":"0.1.0 to 0.1.7","lineage":"DeepNovo (despite an 'adapted from Casanovo' comment)","rule":"0.1 / 0.5 Da, forward only, with the resynchronisation step inverted","order":"correct","empty":"excluded from precision","peptide":"recall: correct / all spectra; precision: correct / non-empty predictions","curve":"peptide precision against recall at 21 confidence quantiles","repo":"instadeepai/instanovo","ref":"43671f0","path":"instanovo/utils/metrics.py","url":"https://github.com/instadeepai/instanovo/blob/43671f0/instanovo/utils/metrics.py"},{"tool":"InstaNovo","versions":"1.0.0 to 1.1.4","lineage":"DeepNovo","rule":"0.1 / 0.5 Da, forward only","order":"correct","empty":"excluded from precision; beam search returns nothing when no beam fits the precursor","peptide":"recall: correct / all spectra; precision: correct / non-empty predictions","curve":"as 0.1, plus recall at a given FDR","repo":"instadeepai/instanovo","ref":"971c367","path":"instanovo/utils/metrics.py","url":"https://github.com/instadeepai/instanovo/blob/971c367/instanovo/utils/metrics.py"},{"tool":"InstaNovo","versions":"1.2.0 and later","lineage":"DeepNovo","rule":"0.1 / 0.5 Da, forward only","order":"correct","empty":"excluded from precision; beam search now keeps beams that miss the precursor","peptide":"recall: correct / all spectra; precision: correct / non-empty predictions","curve":"as 1.0","repo":"instadeepai/instanovo","ref":"66f837d","path":"instanovo/inference/beam_search.py","url":"https://github.com/instadeepai/instanovo/blob/66f837d/instanovo/inference/beam_search.py"},{"tool":"InstaNovo-P","versions":"latest","lineage":"InstaNovo 0.1","rule":"0.1 / 0.5 Da, forward only, with the 0.1 inverted step; PTM-site precision and recall","order":"correct","empty":"excluded from precision","peptide":"as InstaNovo","curve":"as InstaNovo","repo":"InstaDeepAI/InstaNovo-P","ref":"1f61ac0","path":"instanovo/utils/metrics.py","url":"https://github.com/InstaDeepAI/InstaNovo-P/blob/1f61ac0/instanovo/utils/metrics.py"},{"tool":"Winnow","versions":"latest","lineage":"InstaNovo 1.x matcher","rule":"0.1 / 0.5 Da, forward only","order":"FDR only","empty":"n/a","peptide":"used to label correct PSMs for a labelled FDR","curve":"FDR from calibrated confidence","repo":"instadeepai/winnow","ref":"1bde79e","path":"winnow/fdr/nonparametric.py","url":"https://github.com/instadeepai/winnow/blob/1bde79e/winnow/fdr/nonparametric.py"},{"tool":"GraphNovo","versions":"latest","lineage":"DeepNovo","rule":"0.1 / 0.5 Da, forward only, prefix measured at the residue's start","order":"correct","empty":"kept","peptide":"correct / all spectra","curve":"none","repo":"AmadeusloveIris/Graphnovo","ref":"e05651f","path":"genova/task/edge_version_inference.py","url":"https://github.com/AmadeusloveIris/Graphnovo/blob/e05651f/genova/task/edge_version_inference.py"},{"tool":"DIANovo","versions":"latest","lineage":"DeepNovo","rule":"0.1 / 0.5 Da, forward only","order":"correct","empty":"spectra whose label lookup fails leave every denominator","peptide":"correct / evaluated spectra","curve":"none","repo":"hearthewind/dianovo","ref":"cc413f2","path":"main_model/rnova/task/sequence_generation_inference_utils.py","url":"https://github.com/hearthewind/dianovo/blob/cc413f2/main_model/rnova/task/sequence_generation_inference_utils.py"},{"tool":"DiNovo","versions":"latest","lineage":"independent","rule":"residue identity and a 0.05 Da prefix, forward only","order":"correct","empty":"unsupported labels leave every denominator","peptide":"top-k cumulative peptide recall","curve":"none","repo":"YanFuGroup/DiNovo","ref":"9f4bb0d","path":"MSValidationXueli.py","url":"https://github.com/YanFuGroup/DiNovo/blob/9f4bb0d/MSValidationXueli.py"},{"tool":"PowerNovo2","versions":"latest","lineage":"Casanovo (older copy)","rule":"0.1 / 0.5 Da, prefix then suffix","order":"correct","empty":"a file with any empty prediction is silently skipped","peptide":"printed as peptide precision, computed as correct / all rows, per file","curve":"AA precision-recall plot that counts each correct residue as a negative too (precision capped at 0.5)","repo":"protdb/PowerNovo2","ref":"b6afc8a","path":"benchmark/evaluate/plot_pr_curves.py","url":"https://github.com/protdb/PowerNovo2/blob/b6afc8a/benchmark/evaluate/plot_pr_curves.py"},{"tool":"pUniFind","versions":"latest","lineage":"Casanovo (older copy)","rule":"0.1 / 0.5 Da, prefix then suffix, I read as L","order":"swapped","empty":"dropped","peptide":"correct / all spectra","curve":"none","repo":"pFindStudio/pUniFind","ref":"f7b8ad0","path":"PepMS/models/ms_pep.py","url":"https://github.com/pFindStudio/pUniFind/blob/f7b8ad0/PepMS/models/ms_pep.py"},{"tool":"PepGo (benchmark script)","versions":"latest","lineage":"independent","rule":"token identity position by position","order":"recall only","empty":"kept","peptide":"exact string","curve":"none","repo":"alifare/PepGo","ref":"0f0f5cb","path":"tools/benchmark.py","url":"https://github.com/alifare/PepGo/blob/0f0f5cb/tools/benchmark.py"},{"tool":"denovo_benchmarks","versions":"2024-03","lineage":"Casanovo","rule":"0.1 / 0.5 Da, prefix then suffix","order":"correct","empty":"dropped","peptide":"correct / predicted spectra","curve":"none","repo":"bittremieuxlab/denovo_benchmarks","ref":"827d403","path":"evaluate.py","url":"https://github.com/bittremieuxlab/denovo_benchmarks/blob/827d403/evaluate.py"},{"tool":"denovo_benchmarks","versions":"2024-04-12 to 2025-02-07","lineage":"Casanovo","rule":"0.1 / 0.5 Da, prefix then suffix","order":"swapped","empty":"dropped","peptide":"correct / predicted spectra","curve":"precision against coverage","repo":"bittremieuxlab/denovo_benchmarks","ref":"91df68b","path":"evaluate.py","url":"https://github.com/bittremieuxlab/denovo_benchmarks/blob/91df68b/evaluate.py"},{"tool":"denovo_benchmarks","versions":"2025-02-08 to 2026-09-25","lineage":"Casanovo","rule":"0.1 / 0.5 Da, prefix then suffix","order":"swapped","empty":"kept","peptide":"printed as peptide precision, computed as correct / all labelled spectra","curve":"precision against coverage, 200 points","repo":"bittremieuxlab/denovo_benchmarks","ref":"f8efb22","path":"evaluate.py","url":"https://github.com/bittremieuxlab/denovo_benchmarks/blob/f8efb22/evaluate.py"},{"tool":"denovo_benchmarks","versions":"2026-09-26 and later","lineage":"Casanovo","rule":"0.1 / 0.5 Da, prefix then suffix","order":"correct","empty":"kept","peptide":"printed as peptide precision, computed as correct / all labelled spectra","curve":"precision against coverage, 200 points","repo":"bittremieuxlab/denovo_benchmarks","ref":"3a92eba","path":"evaluation/evaluate.py","url":"https://github.com/bittremieuxlab/denovo_benchmarks/blob/3a92eba/evaluation/evaluate.py"},{"tool":"ProteoBench","versions":"current","lineage":"Casanovo, in ppm","rule":"20 ppm per residue and 50 ppm prefix, prefix then suffix; an exact mode adds token identity","order":"correct","empty":"dropped from precision, kept in recall","peptide":"precision and recall reported separately","curve":"precision against coverage at every distinct score (since 2026-08-04)","repo":"Proteobench/ProteoBench","ref":"9a886036","path":"proteobench/score/denovoscores.py","url":"https://github.com/Proteobench/ProteoBench/blob/9a886036/proteobench/score/denovoscores.py"}]})How de novo sequencing is scored
The metrics deep-learning papers report, and how their code computes them
Peptide recall, amino-acid precision, precision against coverage and the rest: what deep-learning de novo sequencing papers say they measured, and what the evaluation code of each tool and each of its versions actually computes. Every number links to the code that computes it.
Nearly every deep-learning de novo sequencing paper reports the same few numbers: peptide recall, amino-acid precision and, more and more often, a curve of precision against coverage. They read like one standard. This page checks whether they are, twice over: against what 115 papers say they computed, and against the evaluation code of 45 implementations, version by version.
How it was done. The papers are the catalog’s 124 deep-learning method papers; 9 of them have no full text in the catalog’s PDF library and are left out. Every metric a paper reports, with its definition quoted and the section it came from, is in papers.jsonl. The code was read at the commit each row of the implementation table links to, on 2026-11-22. Every number is a link to an entry in Reproducing every number; the worked examples run transcriptions of each tool’s matching loop, and the entry names the file and commit each was copied from.
Four things to know before putting two numbers side by side.
- “Peptide precision” is usually a recall. Casanovo’s code calls its peptide figure
pep_precisionand divides it by every spectrum, and NovoBench, denovo_benchmarks and PowerNovo2 do the same. 12 papers define a “peptide precision” over all spectra; 3 define a “peptide recall” over predicted peptides only. - Amino-acid precision and recall are swapped in a whole family of models. Casanovo v3.0 to v3.5 passed the prediction where the truth belonged, and v4.0 fixed it. The models that copied the v3 code did not: 9 entries in the table below still swap the two, among them ContraNovo, PrimeNovo, RefineNovo, RankNovo and π-HelixNovo. denovo_benchmarks swapped them from April 2024 to September 2026.
- “Correct” means the right mass, not the right sequence. The standard rule accepts K for Q (0.036 Da apart), F for oxidised M (0.033 Da) and D for deamidated N (0.000 Da), at peptide level too. ProteoBench’s ppm tolerances and Casanovo-TMT’s exact-peptidoform rule do not.
- “Peptide AUC” names at least three different areas. NovoBench integrates precision against recall, which can never exceed the peptide recall; others integrate against coverage. For the same 1,000 predictions, half of them right and ranked first, the two give 0.50 and 0.85.
What the papers report
97 of the 115 papers report a peptide-level figure, under one name or another, 78 report amino-acid precision or recall, and 54 a precision-coverage curve or the area under it. Fewer report an FDR (25), PTM precision or recall (21) or calibration (9), and 23 introduce a metric of their own.
Where the code disagrees
Nearly every implementation descends from one of two loops: DeepNovo’s, which walks forward only, and Casanovo’s evaluate.py, which adds the pass from the C-terminus. What separates the numbers is mostly not the rule but what happens around it: which argument is the truth, which spectra are counted, and which version of the code ran.
Forward only, or forward and back
DeepNovo’s loop walks both peptides from the N-terminus. When the prefix masses drift apart by more than 0.5 Da, as they do after an inserted or wrong-mass residue, nothing after the drift can match again. Casanovo’s second pass from the C-terminus recovers the residues after the error. Truth PEPTIDEK against the prediction PEPGTIDEK matches 3 residues forward only and 8 with Casanovo’s rule. 12 implementations in the table are forward only: the DeepNovo family, GraphNovo, DIANovo, SMSNet, DiNovo and InstaNovo, whose matcher carries an “adapted from Casanovo” comment but is DeepNovo’s loop. Their amino-acid numbers come out lower on the same predictions; their peptide numbers do not change, because a correct peptide never drifts.
Precision and recall, swapped
Casanovo’s aa_match_batch returns the residue counts in the order its two peptide arguments came in, and aa_match_metrics expects the true count first. From v3.0.0 to v3.5.0 the validation step passed the prediction first, so the number printed as amino-acid precision was the recall and the other way round; commit 3a18fed fixed it for v4.0.0. For truth PEPTIDEK and prediction PEPTIDE, precision is 1.000 and recall 0.875; the swapped code prints 0.875 as precision. Peptide figures are unaffected, because the match itself is symmetric.
Most of the models built on Casanovo v3 kept the swap, and their latest code still has it: ContraNovo, PrimeNovo and every copy of it (π-PrimeNovo, ScopeNovo), RefineNovo, RankNovo, DiffNovo, Transformer-DIA and DiffNovo-DIA, and every tag of π-HelixNovo. TSARseqNovo copied the fixed v4.0 file and kept the old argument order; pUniFind’s own copy swaps them too. XuanjiNovo swapped them at its v1.1 tag and has since fixed it, and π-MNovo fixed it in v0.2.0. The multi-tool benchmark denovo_benchmarks reversed the arguments in April 2024 and fixed it on 26 September 2026 (commit 3a92eba), so its “AA precision” from that period is amino-acid recall.
Which spectra are counted
Up to v3.5.0, Casanovo dropped every spectrum whose prediction was empty or malformed from both sides of the comparison, so its peptide figure was computed over the spectra that got a usable prediction, and a model that abstained more scored higher. ContraNovo, the PrimeNovo family, RefineNovo, DiffNovo and π-HelixNovo inherited that. From v4.0 Casanovo keeps them; NovoBench, AdaNovo and dIon keep them too.
InstaNovo excludes empty predictions from precision only, which is what precision means. But what counts as empty changed: before 1.2.0, beam search returned nothing when no beam fitted the precursor mass, so precision was computed over precursor-matched predictions only. From 1.2.0 those beams are kept, ranked after the matching ones, so InstaNovo precision before and after 1.2.0 is not like-for-like; the recall is. denovo_benchmarks made the opposite move on 8 February 2025: until then its “Pep precision” was correct over predicted spectra, a precision, and from then on over all labelled spectra, a recall.
One tool, several versions
Casanovo.
- v1.1 to v2.1 matched on the mass before each residue in a single pass.
- v3.0 rewrote that as the prefix-then-suffix rule used since.
- v3.0 to v3.5 swapped precision and recall and dropped empty predictions. v4.0 fixed both, but in validation a spectrum whose beams were all discarded silently shifts the rest of its batch onto the wrong truth.
- v5.0 to v5.1.2:
casanovo sequence --evaluatecounted the padding of the true sequence as residues, so it printed a peptide precision of 0. v5.2.0 fixed it. - v5: the validation code still splits a peptide before every capital letter, so
M[Oxidation]becomes two pieces of mass 0, and any modified residue matches any other. - The confidence that orders a precision-coverage curve changed three times: the mean of the residue scores minus 1 off-precursor (v3 to v4.3), the product minus 1 (v5.0 to v5.1), and the product with no penalty (v5.2 on). The same model gives three different curves.
InstaNovo.
- 0.1.0 to 0.1.7 had the resynchronisation step inverted: after the first mismatch it advanced the peptide that was already ahead, so no later residue could match. GGPEPTIDE against NPEPTIDE matches 0 residues in 0.1.x and 7 from 1.0.0. Peptide figures were unaffected. InstaNovo-P still ships the 0.1 step, including in its PTM matcher.
- 1.0.0 made its amino-acid error rate count I and L as equal; 0.1 counted every I/L call as an error.
- 1.2.0 changed which predictions are empty, as above.
denovo_benchmarks has had four evaluators: correct but over predicted spectra only (March 2024), swapped (from April 2024), swapped and over all labelled spectra (from February 2025), and correct over all labelled spectra (from September 2026). ProteoBench stored no AUC until 4 August 2026, and has since integrated precision against coverage at every distinct score.
Smaller differences that still change a number
- DeepNovo’s peptide rule ignores the prediction’s length: a peptide counts as correct when every true residue is matched, so PEPTIDEK is a correct call for PEPTIDE. Its DIA and PointNovo descendants, given more than one candidate per spectrum, keep the candidate closest to the truth; with the default of one it makes no difference.
- DiffNovo logs, per batch, the better of its two decoders, which no deployed system could choose, and its peptide slot holds an amino-acid value.
- RankNovo averages per-batch peptide recalls without weighting them by batch size, and weights the amino-acid figure by the first candidate’s length.
- MemNovo strips every modification before matching and its mass table has no glycine, so a deleted G costs nothing; under its PrimeNovo model name it falls back to counting letters regardless of order.
- PowerNovo2’s precision-recall plot counts each correct residue as a negative as well, which caps the plotted precision at 0.5, and a results file holding any empty prediction is silently skipped.
- PTM precision and recall in NovoBench and AdaNovo inherit the mass rule, so F in place of an oxidised M counts toward PTM recall.
- Unknown tokens weigh 0 in Casanovo and its copies, so a label with an unparsed modification is scored with every later prefix mass off; InstaNovo and the DeepNovo family refuse it.
Three areas called AUC
- NovoBench (and LIPNovo and SemiNovo, which copy it) integrates precision against recall: the area can never exceed the peptide recall, and wrong predictions add no width.
- Precision against coverage, coverage being the share of all spectra predicted, is what denovo_benchmarks (at 200 downsampled points), ProteoBench (at every distinct score), dIon and Casanovo’s unreleased
devbranch integrate. The Pairwise scripts compute coverage over predicted spectra only. - InstaNovo integrates peptide precision against recall at 21 confidence quantiles.
On the same predictions the first two give 0.50 and 0.85; neither is wrong, but they are not the same number.
Metrics beyond recall
RankNovo’s PMD and RMD
RankNovo reranks the candidates of several sequencers, and to train the reranker it introduced two measures of how far a candidate is from the truth. Both are training targets, not evaluation metrics: the paper still reports peptide recall and amino-acid precision, and the released code computes those with the swapped Casanovo v3 call.
PMD, Peptide Mass Deviation, is an edit distance in which inserting or deleting a residue costs one unit and substituting one costs the mass difference; the released code fixes the unit at 51.4238 Da. It is not the difference between the two peptides’ masses: GG against N, which weigh the same, costs 2.11 units, a missing residue costs 1 unit whatever it weighs, and K for Q costs 0.0007.
RMD, Residual Mass Deviation, gives each residue of the candidate the signed offset between its prefix mass and the nearest prefix mass of the truth. For PEPGGK against PEPNK it is -57.02 Da at the first G and 0.00 Da at the end, where the masses agree again. In the released inference code only the PMD head chooses the candidate; the RMD head is computed and discarded. No other repository implements either.
Error rates without a search engine
A recall needs labels, which only database search provides, so the field has started to estimate error rates on unlabelled spectra: Winnow turns calibrated confidence into an FDR (the mean of one minus the confidence above a threshold); DiNovo maps predictions to a target-decoy protein database; π-NovoQC compares counterpart spectra; Casanovo’s 2023 preprint and Casanovo-TMT count matches to a shuffled or decoy digest, the latter saying plainly that this is not a search-engine FDR; and Cascadia counts a DIA prediction correct when it matches the reference proteome exactly.
Other measures
Casanovo 5 reports a calibration curve. DiffuNovo reports the error of the predicted peptide’s mass and the share within 0.001 Da. BiATNovo adds a position-weighted BLEU and a mass-aware alignment score, the QuasiNovo thesis a longest common subsequence anchored on terminal masses, OmniNovo backbone and modification-site accuracy, Spectralis precision and recall of its fragment-bin corrections, and NovoCRF recall split by how much fragment-ion evidence each spectrum holds.
Reading a number
Before comparing two published figures, find out:
- Which code, at which version, computed it. A Casanovo v3 descendant and a Casanovo v4 one print the same labels for different quantities.
- The denominator. All labelled spectra, spectra with a prediction, or spectra with a well-formed prediction.
- The residue rule. The 0.1 / 0.5 Da mass rule, ppm, or an exact sequence.
- For a curve, which confidence orders it, and what the area is taken against.
Reproducing every number
Every number above, in the order it first appears: what it counts, the public files it comes from, the scripts that produce its input (in run_all.sh’s order), and the code in page_numbers.py that computes it. page_numbers.py --only <id> recomputes one.
115 (Of those, papers whose full text was read). Reads papers.jsonl: every metric each paper reports, with its definition quoted. Input from papers.jsonl. Recompute: page_numbers.py --only n_read.
Code
@number("n_read", "Of those, papers whose full text was read", ["papers"], ["papers.jsonl"])
def _():
return len(papers())45 (Implementations compared (a tool, or one version range of it)). Reads implementations.json: one row per tool and version range, with the commit read. Input from implementations.json. Recompute: page_numbers.py --only n_impl.
Code
@number("n_impl", "Implementations compared (a tool, or one version range of it)", ["code"], ["implementations.json"])
def _():
return len(code())124 (Deep-learning method papers in the catalog considered here). Reads papers.jsonl: every metric each paper reports, with its definition quoted. Input from papers.jsonl. Recompute: page_numbers.py --only n_papers.
Code
@number("n_papers", "Deep-learning method papers in the catalog considered here", ["papers"], ["papers.jsonl"])
def _():
return len(all_papers())9 (Papers with no full text in the library (paywalled or not yet fetched)). Reads papers.jsonl: every metric each paper reports, with its definition quoted. Input from papers.jsonl. Recompute: page_numbers.py --only n_unread.
Code
@number("n_unread", "Papers with no full text in the library (paywalled or not yet fetched)", ["papers"], ["papers.jsonl"])
def _():
return len(all_papers()) - len(papers())12 (Papers that call a peptide figure ‘precision’ and define it over all spectra). Reads papers.jsonl: every metric each paper reports, with its definition quoted. Input from papers.jsonl, build_site_data.py. Recompute: page_numbers.py --only n_prec_named_recall.
Code
@number("n_prec_named_recall", "Papers that call a peptide figure 'precision' and define it over all spectra", ["papers"], ["papers.jsonl", "build_site_data.py"])
def _():
return sum(p["naming"] == "precision over all spectra" for p in site()["papers"])3 (Papers that call a peptide figure ‘recall’ and define it over predicted peptides). Reads papers.jsonl: every metric each paper reports, with its definition quoted. Input from papers.jsonl, build_site_data.py. Recompute: page_numbers.py --only n_recall_named_prec.
Code
@number("n_recall_named_prec", "Papers that call a peptide figure 'recall' and define it over predicted peptides", ["papers"], ["papers.jsonl", "build_site_data.py"])
def _():
return sum(p["naming"] == "recall over predictions" for p in site()["papers"])9 (Entries in the table whose latest code still swaps amino-acid precision and recall). Reads implementations.json: one row per tool and version range, with the commit read. Input from implementations.json. Recompute: page_numbers.py --only n_swapped_now.
Code
@number("n_swapped_now", "Entries in the table whose latest code still swaps amino-acid precision and recall", ["code"], ["implementations.json"])
def _():
latest = {}
for c in code():
latest[c["tool"]] = c # rows are in version order, so the last one per tool is the latest
return sum(c["order"] == "swapped" for c in latest.values())0.036 (Mass difference between K and Q, in Da). Reads Casanovo v4.0.0 residue masses, casanovo/config.yaml. Recompute: page_numbers.py --only d_kq.
Code
@number("d_kq", "Mass difference between K and Q, in Da", ["casanovo_masses"], fmt="{:.3f}")
def _():
return _da("K", "Q")0.033 (Mass difference between F and oxidised M, in Da). Reads Casanovo v4.0.0 residue masses, casanovo/config.yaml. Recompute: page_numbers.py --only d_fm.
Code
@number("d_fm", "Mass difference between F and oxidised M, in Da", ["casanovo_masses"], fmt="{:.3f}")
def _():
return _da("F", "m")0.000 (Mass difference between deamidated N and D, in Da). Reads Casanovo v4.0.0 residue masses, casanovo/config.yaml. Recompute: page_numbers.py --only d_nd.
Code
@number("d_nd", "Mass difference between deamidated N and D, in Da", ["casanovo_masses"], fmt="{:.3f}")
def _():
return _da("n", "D")0.50 (NovoBench’s AUC for 1,000 spectra, half right and all of those ranked first). Reads NovoBench, novobench/metrics/evaluate.py (curve_auc). Recompute: page_numbers.py --only ex_auc_half_nb.
Code
@number("ex_auc_half_nb", "NovoBench's AUC for 1,000 spectra, half right and all of those ranked first", ["novobench"], fmt="{:.2f}")
def _():
return novobench_auc(HALF)0.85 (Area under precision against coverage for the same 1,000 spectra (trapezoid, coverage = rank / N)). Reads NovoBench, novobench/metrics/evaluate.py (curve_auc). Recompute: page_numbers.py --only ex_auc_half_cov.
Code
@number("ex_auc_half_cov", "Area under precision against coverage for the same 1,000 spectra (trapezoid, coverage = rank / N)", ["novobench"], fmt="{:.2f}")
def _():
n = len(HALF)
cs = list(accumulate(int(c) for c in HALF))
prec = [c / (k + 1) for k, c in enumerate(cs)]
cov = [(k + 1) / n for k in range(n)]
return sum((cov[k] - cov[k - 1]) * (prec[k] + prec[k - 1]) / 2 for k in range(1, n))97 (Papers reporting a peptide-level figure (recall or precision, under any name)). Reads papers.jsonl: every metric each paper reports, with its definition quoted. Input from papers.jsonl. Recompute: page_numbers.py --only n_pep.
Code
@number("n_pep", "Papers reporting a peptide-level figure (recall or precision, under any name)", ["papers"], ["papers.jsonl"])
def _():
return sum(reports(r, "peptide_recall", "peptide_precision") for r in papers())78 (Papers reporting amino-acid precision or recall). Reads papers.jsonl: every metric each paper reports, with its definition quoted. Input from papers.jsonl. Recompute: page_numbers.py --only n_aa.
Code
@number("n_aa", "Papers reporting amino-acid precision or recall", ["papers"], ["papers.jsonl"])
def _():
return sum(reports(r, "aa_precision", "aa_recall") for r in papers())54 (Papers reporting a precision-coverage curve or its area). Reads papers.jsonl: every metric each paper reports, with its definition quoted. Input from papers.jsonl. Recompute: page_numbers.py --only n_curve.
Code
@number("n_curve", "Papers reporting a precision-coverage curve or its area", ["papers"], ["papers.jsonl"])
def _():
return sum(reports(r, "precision_coverage_curve", "AUC_of_precision_coverage") for r in papers())25 (Papers reporting an FDR-based figure). Reads papers.jsonl: every metric each paper reports, with its definition quoted. Input from papers.jsonl. Recompute: page_numbers.py --only n_fdr.
Code
@number("n_fdr", "Papers reporting an FDR-based figure", ["papers"], ["papers.jsonl"])
def _():
return sum(reports(r, "FDR-based") for r in papers())21 (Papers reporting PTM precision or recall). Reads papers.jsonl: every metric each paper reports, with its definition quoted. Input from papers.jsonl. Recompute: page_numbers.py --only n_ptm.
Code
@number("n_ptm", "Papers reporting PTM precision or recall", ["papers"], ["papers.jsonl"])
def _():
return sum(reports(r, "PTM_precision/recall", "PTM_precision") for r in papers())9 (Papers reporting a calibration metric). Reads papers.jsonl: every metric each paper reports, with its definition quoted. Input from papers.jsonl. Recompute: page_numbers.py --only n_calib.
Code
@number("n_calib", "Papers reporting a calibration metric", ["papers"], ["papers.jsonl"])
def _():
return sum(reports(r, "calibration metric") for r in papers())23 (Papers that introduce a metric of their own). Reads papers.jsonl: every metric each paper reports, with its definition quoted. Input from papers.jsonl. Recompute: page_numbers.py --only n_new_metric.
Code
@number("n_new_metric", "Papers that introduce a metric of their own", ["papers"], ["papers.jsonl"])
def _():
return sum(bool(r.get("new_metrics")) for r in papers())1 (Truth PEPTIDEK, prediction PEPTIDEQ: does the 0.1 / 0.5 Da rule count the peptide correct? (1 = yes)). Reads Casanovo v5.2.1, casanovo/denovo/evaluate.py (aa_match_prefix, aa_match_prefix_suffix). Recompute: page_numbers.py --only ex_kq_pep.
Code
@number("ex_kq_pep", "Truth PEPTIDEK, prediction PEPTIDEQ: does the 0.1 / 0.5 Da rule count the peptide correct? (1 = yes)", ["casanovo_eval"])
def _():
return int(all(casanovo_best("PEPTIDEK", "PEPTIDEQ")))284 (K against Q in ppm of K’s mass (ProteoBench allows 20 ppm per residue)). Reads Casanovo v4.0.0 residue masses, casanovo/config.yaml, ProteoBench, proteobench/score/denovoscores.py (20 and 50 ppm). Recompute: page_numbers.py --only ppm_kq.
Code
@number("ppm_kq", "K against Q in ppm of K's mass (ProteoBench allows 20 ppm per residue)", ["casanovo_masses", "proteobench"], fmt="{:,.0f}")
def _():
return _da("K", "Q") / MASS["K"] * 1e643 (Papers that state the 0.1 Da residue / 0.5 Da prefix rule). Reads papers.jsonl: every metric each paper reports, with its definition quoted. Input from papers.jsonl, build_site_data.py. Recompute: page_numbers.py --only n_rule_mass.
Code
@number("n_rule_mass", "Papers that state the 0.1 Da residue / 0.5 Da prefix rule", ["papers"], ["papers.jsonl", "build_site_data.py"])
def _():
return _rule_count("0.1 / 0.5 Da mass rule")17 (Papers that count a residue or peptide correct only on an exact sequence match). Reads papers.jsonl: every metric each paper reports, with its definition quoted. Input from papers.jsonl, build_site_data.py. Recompute: page_numbers.py --only n_rule_exact.
Code
@number("n_rule_exact", "Papers that count a residue or peptide correct only on an exact sequence match", ["papers"], ["papers.jsonl", "build_site_data.py"])
def _():
return _rule_count("Exact sequence")4 (Papers with a stricter mass rule or position-by-position matching). Reads papers.jsonl: every metric each paper reports, with its definition quoted. Input from papers.jsonl, build_site_data.py. Recompute: page_numbers.py --only n_rule_other.
Code
@number("n_rule_other", "Papers with a stricter mass rule or position-by-position matching", ["papers"], ["papers.jsonl", "build_site_data.py"])
def _():
return _rule_count("Stricter mass rule") + _rule_count("Position by position")51 (Papers that state no matching rule at all). Reads papers.jsonl: every metric each paper reports, with its definition quoted. Input from papers.jsonl, build_site_data.py. Recompute: page_numbers.py --only n_rule_none.
Code
@number("n_rule_none", "Papers that state no matching rule at all", ["papers"], ["papers.jsonl", "build_site_data.py"])
def _():
return _rule_count("Not stated")24 (Papers that say I and L are counted as equal). Reads papers.jsonl: every metric each paper reports, with its definition quoted. Input from papers.jsonl, build_site_data.py. Recompute: page_numbers.py --only n_il_equal.
Code
@number("n_il_equal", "Papers that say I and L are counted as equal", ["papers"], ["papers.jsonl", "build_site_data.py"])
def _():
return sum(p["il"] == "Equal" for p in site()["papers"] if p["read"])88 (Papers that do not say how I and L are counted). Reads papers.jsonl: every metric each paper reports, with its definition quoted. Input from papers.jsonl, build_site_data.py. Recompute: page_numbers.py --only n_il_unstated.
Code
@number("n_il_unstated", "Papers that do not say how I and L are counted", ["papers"], ["papers.jsonl", "build_site_data.py"])
def _():
return sum(p["il"] == "Not stated" for p in site()["papers"] if p["read"])3 (PEPTIDEK against PEPGTIDEK: residues matched forward only (DeepNovo, InstaNovo 1.x)). Reads DeepNovo, deepnovo_worker_test.py (_match_AA_novor), InstaNovo 1.0.0, instanovo/utils/metrics.py (_novor_match). Recompute: page_numbers.py --only ex_fwd.
Code
@number("ex_fwd", "PEPTIDEK against PEPGTIDEK: residues matched forward only (DeepNovo, InstaNovo 1.x)", ["deepnovo", "instanovo_100"])
def _():
return forward_only("PEPTIDEK", "PEPGTIDEK")8 (PEPTIDEK against PEPGTIDEK: residues matched with Casanovo’s prefix-then-suffix rule). Reads Casanovo v5.2.1, casanovo/denovo/evaluate.py (aa_match_prefix, aa_match_prefix_suffix). Recompute: page_numbers.py --only ex_best.
Code
@number("ex_best", "PEPTIDEK against PEPGTIDEK: residues matched with Casanovo's prefix-then-suffix rule", ["casanovo_eval"])
def _():
return sum(casanovo_best("PEPTIDEK", "PEPGTIDEK"))12 (Implementations that match residues in a forward pass only, with no suffix pass). Reads implementations.json: one row per tool and version range, with the commit read. Input from implementations.json. Recompute: page_numbers.py --only n_forward_only.
Code
@number("n_forward_only", "Implementations that match residues in a forward pass only, with no suffix pass", ["code"], ["implementations.json"])
def _():
return sum("forward only" in c["rule"] for c in code())1.000 (Truth PEPTIDEK, prediction PEPTIDE: amino-acid precision, correctly computed). Reads Casanovo v5.2.1, casanovo/denovo/evaluate.py (aa_match_prefix, aa_match_prefix_suffix). Recompute: page_numbers.py --only ex_swap_prec.
Code
@number("ex_swap_prec", "Truth PEPTIDEK, prediction PEPTIDE: amino-acid precision, correctly computed", ["casanovo_eval"], fmt="{:.3f}")
def _():
return sum(casanovo_best("PEPTIDEK", "PEPTIDE")) / len("PEPTIDE")0.875 (Truth PEPTIDEK, prediction PEPTIDE: amino-acid recall, which swapped code prints as precision). Reads Casanovo v5.2.1, casanovo/denovo/evaluate.py (aa_match_prefix, aa_match_prefix_suffix), Casanovo v3.5.0, casanovo/denovo/model.py (validation_step). Recompute: page_numbers.py --only ex_swap_rec.
Code
@number("ex_swap_rec", "Truth PEPTIDEK, prediction PEPTIDE: amino-acid recall, which swapped code prints as precision", ["casanovo_eval", "casanovo_v35"], fmt="{:.3f}")
def _():
return sum(casanovo_best("PEPTIDEK", "PEPTIDE")) / len("PEPTIDEK")0 (GGPEPTIDE against NPEPTIDE: residues matched by InstaNovo 0.1.x). Reads InstaNovo 0.1.0, instanovo/utils/metrics.py (_novor_match). Recompute: page_numbers.py --only ex_inv_01.
Code
@number("ex_inv_01", "GGPEPTIDE against NPEPTIDE: residues matched by InstaNovo 0.1.x", ["instanovo_010"])
def _():
return forward_only("GGPEPTIDE", "NPEPTIDE", inverted=True)7 (GGPEPTIDE against NPEPTIDE: residues matched by InstaNovo 1.0 and later). Reads InstaNovo 1.0.0, instanovo/utils/metrics.py (_novor_match). Recompute: page_numbers.py --only ex_inv_10.
Code
@number("ex_inv_10", "GGPEPTIDE against NPEPTIDE: residues matched by InstaNovo 1.0 and later", ["instanovo_100"])
def _():
return forward_only("GGPEPTIDE", "NPEPTIDE")2.11 (RankNovo PMD between GG and N, which have the same total mass (units of 51.4238 Da)). Reads RankNovo, RankNovo/RankNovo/msa/tokenizer.py (TMmass, Greedymass). Recompute: page_numbers.py --only pmd_gg_n.
Code
@number("pmd_gg_n", "RankNovo PMD between GG and N, which have the same total mass (units of 51.4238 Da)", ["ranknovo"], fmt="{:.2f}")
def _():
return ranknovo_pmd("PEPGGK", "PEPNK")1 (RankNovo PMD for a missing residue, whatever its mass (units of 51.4238 Da)). Reads RankNovo, RankNovo/RankNovo/msa/tokenizer.py (TMmass, Greedymass). Recompute: page_numbers.py --only pmd_del.
Code
@number("pmd_del", "RankNovo PMD for a missing residue, whatever its mass (units of 51.4238 Da)", ["ranknovo"], fmt="{:.0f}")
def _():
return ranknovo_pmd("PEPTDE", "PEPTIDE")0.0007 (RankNovo PMD between PEPTIDEK and PEPTIDEQ (units of 51.4238 Da)). Reads RankNovo, RankNovo/RankNovo/msa/tokenizer.py (TMmass, Greedymass). Recompute: page_numbers.py --only pmd_kq.
Code
@number("pmd_kq", "RankNovo PMD between PEPTIDEK and PEPTIDEQ (units of 51.4238 Da)", ["ranknovo"], fmt="{:.4f}")
def _():
return ranknovo_pmd("PEPTIDEK", "PEPTIDEQ")-57.02 (RankNovo RMD at the first G of PEPGGK against PEPNK, in Da). Reads RankNovo, RankNovo/RankNovo/msa/tokenizer.py (TMmass, Greedymass). Recompute: page_numbers.py --only rmd_gg_mid.
Code
@number("rmd_gg_mid", "RankNovo RMD at the first G of PEPGGK against PEPNK, in Da", ["ranknovo"], fmt="{:.2f}")
def _():
return ranknovo_rmd("PEPGGK", "PEPNK")[3] * GAP0.00 (RankNovo RMD at the last residue of PEPGGK against PEPNK, in Da (0: the masses have realigned)). Reads RankNovo, RankNovo/RankNovo/msa/tokenizer.py (TMmass, Greedymass). Recompute: page_numbers.py --only rmd_gg_last.
Code
@number("rmd_gg_last", "RankNovo RMD at the last residue of PEPGGK against PEPNK, in Da (0: the masses have realigned)", ["ranknovo"], fmt="{:.2f}")
def _():
return abs(ranknovo_rmd("PEPGGK", "PEPNK")[-1] * GAP)