GraphNovo

algorithm · GNN

GraphNovo: algorithm · GNN. Missing fragments

Missing fragments

Kind	algorithm
Deep learning	yes
Acquisition	DDA
Family	GNN

Code

• https://github.com/AmadeusloveIris/Graphnovo

Live stars, open issues and last-push figures are on the Code activity chart.

Checkpoints

Version	Trained on	Host	Licence	Size	Checked
dataset and checkpoint	—	Zenodo archival	CC-BY-4.0	2.4 GB	live 2026-10-02
original checkpoint	—	Zenodo archival	CC-BY-4.0	2.4 GB	live 2026-10-02

Data

What the papers describing this method trained it on and reported its results on, as each paper states it. A version is named only where the paper names one.

Trained on

• BoxCar acquisition method enables single shot proteomics at a depth of 10,000 proteins in 100 minutes (as deposited)

Evaluated on

• GraphNovo dataset and checkpoint (as deposited)
• NFYB-1 regulates mitochondrial function and longevity via lysosomal prosaposin (as deposited)
• Predictive signatures of 19 antibiotics-induced Escherichia coli proteomes (as deposited)
• The beta subunit of nascent polypeptide associated complex plays a role in flowers and siliques development of Arabidops (as deposited)

Compared against in other papers (2)

2 tables from 2 papers that report this method as a baseline, newest first. The caveats of the comparisons above hold twice over: each paper ran, retrained or quoted this method its own way, so its number here is that paper’s claim, not a measurement of the method. Where a paper says which it did, the word follows the method’s name.

LIPNovo+: Self-Reflective Latent Imputation for Robust De Novo Peptide Sequencing (2026) · 1 table

Table5

LIPNovo+: Self-Reflective Latent Imputation for Robust De Novo Peptide Sequencing, page 18: Comparison with state-of-the-art methods on GraphNovo dataset [11].

GraphNovo dataset and checkpoint	Peptide recall	Peptide AUC	Amino acid recall	Amino acid precision
PEAKS	0.475		0.758	0.775
Novor	0.431		0.740	0.753
DeepNovo	0.497		0.769	0.771
PointNovo	0.601		0.815	0.816
Casanovo · retrained	0.605		0.684	0.684
GraphNovo	0.712		0.876	0.874
LIPNovo	0.729	0.707	0.876	0.882
LIPNovo+	0.740	0.720	0.889	0.891

Latent Imputation before Prediction: A New Computational Paradigm for De Novo Peptide Sequencing (2025) · 1 table

Table4

Latent Imputation before Prediction: A New Computational Paradigm for De Novo Peptide Sequencing, page 7: Comparison with GraphNovo (Mao et al., 2023).

The same table is also printed in Latent Imputation before Prediction: A New Computational Paradigm for De Novo Peptide Sequencing (Table4, page 7).

GraphNovo dataset and checkpoint	Peptide recall	Peptide AUC	Amino acid recall	Amino acid precision
GraphNovo · quoted	0.712		0.876	0.874
LIPNovo	0.729	0.707	0.876	0.882

In the paper: column peptide AUC: the original table bolded LIPNovo (0.707), the only value, since GraphNovo was not run; with nothing to rank it is not marked here.
In the paper: column amino acid recall: the original table only bolded LIPNovo (0.876); GraphNovo (0.876) ties with it and is bolded here too.

Paper describing it

• Mitigating the missing-fragmentation problem in de novo peptide sequencing with a two-stage graph-based deep learning model (2023, Nature Machine Intelligence, peer-reviewed)

Authors (4)

Zeping Mao, Ruixue Zhang, Lei Xin, Ming Li

Seen in the charts

• The architectures
• Code activity
• Models and the authors behind them

Back to the full map