DiffNovo

algorithm · Transformer (NAR)

DiffNovo: algorithm · Transformer (NAR). Transformer-diffusion model

Transformer-diffusion model

Kind algorithm
Deep learning yes
Acquisition DDA
Family Transformer (NAR)

Code

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

Checkpoints

Version Trained on Host Licence Size Checked
pretrained models — Zenodo archival Apache-2.0 456 MB live 2026-10-02

Reported comparisons (2)

The comparison tables this method’s own papers print, standardised: every value on a 0-1 scale, methods down the side, the measure and then the species across. These are numbers papers report about themselves and their baselines. They are not a leaderboard, and they do not compare across tables: each was produced by a different group, on the dataset named in its corner, with each baseline either retrained, run from released weights or quoted from another paper. Where the paper says which, it follows the method’s name (hover it for the sentence); most papers do not say. Bold is the best value in a column and underline the runner-up, our ranking rather than the paper’s own marks.

Table 1

DiffNovo: A Transformer-Diffusion Model for De Novo Peptide Sequencing, page 7: Comparison of DiffNovo and other state-of-the-art (SOTA) methods on amino acid recall and precision across three DIA datasets.

De novo sequencing of DIA data Amino acid recall Amino acid precision
UTI OC Plasma UTI OC Plasma
DiffNovo 0.665 0.749 0.663 0.675 0.743 0.675
DeepNovo-DIA 0.566 0.547 0.689 0.612 0.601 0.702
PepNet 0.396 0.455 0.509 0.391 0.451 0.491
Cascadia 0.293 0.286 0.704 0.341 0.349 0.782

PepNet, Plasma: the printed cell reads “0.491 / 0.725*”, and only 0.491 is in the table. The paper’s note: “* Indicates the positional accuracy reported in [10].”

Table 2

DiffNovo: A Transformer-Diffusion Model for De Novo Peptide Sequencing, page 8: Comparison of DiffNovo and other state-of-the-art (SOTA) methods on peptide precision across three DIA datasets.

De novo sequencing of DIA data Peptide precision
UTI OC Plasma
DiffNovo 0.659 0.735 0.632
DeepNovo-DIA 0.403 0.417 0.351
PepNet 0.296 0.368 0.550
Cascadia 0.291 0.301 0.736

PepNet, Plasma: the printed cell reads “0.550/0.530*/0.664+”, and only 0.550 is in the table. The paper’s note: “+ Indicates the filtered peptide-level accuracy, and * indicates the peptide-level accuracy reported in [10].”

Paper describing it

Authors (3)

Shiva Ebrahimi, Jiancheng Li, Xuan Guo

Seen in the charts

Back to the full map

Back to top