SemiNovo
algorithm · Transformer (AR)
SemiNovo: algorithm · Transformer (AR). Semi-supervised training on identified and 4.5 million unlabeled spectra (DarkSpec), pseudo-labelled by stepwise mass-constrained autoregressive decoding with cumulative…
Semi-supervised training on identified and 4.5 million unlabeled spectra (DarkSpec), pseudo-labelled by stepwise mass-constrained autoregressive decoding with cumulative path confidence, plus head-wise chemical-evidence gating of spectral attention.
| Kind | algorithm |
| Deep learning | yes |
| Acquisition | DDA |
| Family | Transformer (AR) |
Code
Not on the Code activity chart: the GitHub API has returned no figures for this repository, either because it was added after the last daily refresh or because it is private, renamed or deleted.
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
- DarkSpec (v1)
- Nine-species benchmark (InstaNovo split)
- ProteomeTools (HC-PT (NovoBench))
- Seven-species benchmark (NovoBench split)
Evaluated on
- Nine-species benchmark (InstaNovo split)
- ProteomeTools (HC-PT (NovoBench))
- Seven-species benchmark (NovoBench split)
Deposited
- DarkSpec (v1)
Paper describing it
- SemiNovo: Learning Beyond Search-Identified Spectra for De Novo Peptide Sequencing (2026, arXiv, preprint)