Pith. sign in

REVIEW 4 major objections 5 minor 27 references

pUniFind: a unified large pre-trained deep learning model pushing the limit of mass spectra interpretation

T0 review · 4 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read pUniFind unifies database search scoring and open de novo sequencing in one pretrained model trained on over 100 million spectra.

desk verdict A substantive advance in unified MS/MS interpretation, but the headline de novo gains are anchored to the same Open-pFind labels used for training, so they need independent verification. read the letter →

arxiv 2507.00087 v1 pith:73PXFCYL submitted 2025-06-30 cs.LG cs.AI

Jiale Zhao , Pengzhi Mao , Kaifei Wang , Yiming Li , Yaping Peng , Ranfei Chen , Shuqi Lu , Xiaohong Ji
show 7 more authors
Jiaxiang Ding Xin Zhang Yucheng Liao Weinan E Weijie Zhang Han Wen Hao Chi
This is my paper · ORCID
classification cs.LGcs.AI
keywords massspectrometrypeptide-spectrummatchingdenovosequencingopensearchmultimodalpre-trainingproteomicsdeeplearningcross-modalityprediction
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

pUniFind is a large pre-trained deep learning model that treats peptide identification and de novo peptide sequencing as a single multimodal problem, aligning spectra and peptide sequences through cross-modality prediction. The paper claims that, trained on over 100 million spectra annotated by open database search, it rescores database search candidates and performs open, zero-shot de novo sequencing over more than 1,300 modification types, identifying 60% more peptide-spectrum matches than existing de novo methods despite a 300-fold larger search space and 42.6% more peptides in immunopeptidomics. If correct, this would mean a single model can replace handcrafted, feature-based scoring and eliminate the need to prespecify modifications for modified-peptide sequencing. The paper also reports a deep-learning quality-control filter that recovers additional peptides, including 1,891 mapped in the human genome but absent from reference proteomes.

What carries the argument

The machinery is cross-modality pre-training plus a joint scoring head. Separate encoders embed spectra and peptides; pre-training tasks include predicting the spectrum from a peptide, predicting the peptide length, amino-acid count and ion type (b/y, neutral losses) from each peak, and listwise candidate ranking. A joint modality scorer then produces the PSM score used to rerank Open-pFind's top-k candidates, with target-decoy analysis for FDR control. For de novo sequencing, a Peptide Length Aware module predicts a length within ±2 amino acids and generates the sequence token by token; a deep-learning feature filter with predicted spectra and retention times removes unreliable results, and for modification-enriched data a pFind search rescoring step is added.

What would settle it

Re-annotate or synthesize a benchmark where the true peptide sequence is known—for example, synthetic peptides or metabolic-labeling pairs—and compare pUniFind's de novo sequences and modification calls against that ground truth; if the 60% PSM gain and the 42.6% immunopeptidomics gain over baselines shrink or disappear under independent ground truth, the central claim of unified superiority is falsified.

Watch

Extended reading notes

Core claim

The central claim is that end-to-end deep learning, not feature engineering, is the right scoring framework for mass-spectrometry interpretation, and that database search and de novo sequencing share one underlying representation. pUniFind is presented as the first large-scale multimodal pre-trained model to integrate both tasks: it reranks Open-pFind candidate lists with a joint modality scorer and, in the same model, generates peptide sequences with modifications without being told which modifications to expect. The authors report consistent gains over existing engines across nine species, timsTOF and Astral instruments, metaproteomics, and immunopeptidomics, with accuracy checks through entrapment databases, metabolic labeling, and mixed-species searches. The reported 42.6% increase in immunopeptidomic peptide identifications and the 60% increase over de novo baselines are the concrete quantitative claims that carry the argument.

Load-bearing premise

The load-bearing premise is that the Open-pFind annotations used to build the 100-million-spectrum training set are correct enough to serve as ground truth, and that target-decoy FDR estimation remains valid for pUniFind's scores; if the model learns the annotator's errors, the reported gains over Open-pFind and the de novo recall numbers measured against Open-pFind labels would be inflated.

Editorial extensions

If this is right

  • Database search can drop handcrafted scoring features: rescoring candidates with the pretrained model improves peptide identifications, especially when open search expands the candidate space.
  • Modified-peptide de novo sequencing becomes feasible without prespecified modification lists, since modification types are treated as tokens the model can emit.
  • Hard applications benefit most: immunopeptidomics gains 42.6% more peptides than Open-pFind, and metaproteomics and non-tryptic searches also improve.
  • The model transfers to new instrument types with a single epoch of fine-tuning, as shown on timsTOF and Astral data.
  • A deep-learning QC filter can recover additional genome-derived peptides outside reference proteomes while keeping ion coverage, extending the reach of discovery proteomics.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Because the training labels come from Open-pFind, the model may have learned Open-pFind's systematic blind spots; the reported de novo recall against Open-pFind labels could look better than against independent ground truth. That is my inference, not a claim in the paper.
  • The tokenized modification representation treats modification type without site; extending it with site prediction would likely improve site-level accuracy, which the paper already measures and which remains the hardest level.
  • The paper states that retention time is not used and DIA data are not fully exploited; adding either to the same cross-modal framework is the most direct next step and could widen the model's advantage on timsTOF and DIA datasets.
  • The same architecture could generalize to other molecule-spectrum matching problems, such as metabolomics or glycomics, where a spectrum must be aligned to a structured sequence with modifications.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 5 minor

Summary. The manuscript introduces pUniFind, a large pre-trained multimodal deep learning model that jointly performs peptide–spectrum match (PSM) rescoring for database search and open, zero-shot de novo sequencing. The model is trained on over 100 million spectra annotated by Open-pFind using open database search, and it is evaluated on nine-species datasets, Astral and timsTOF data, immunopeptidomics, metaproteomics, and the 21PTM dataset. The central claims are that pUniFind improves peptide identifications by up to 42.6% in immunopeptidomics relative to Open-pFind, identifies 60% more PSMs than existing de novo methods in a 300-fold larger search space, and enables modification-aware de novo sequencing with a peptide-level recall of 63.8% on the 21PTM benchmark. The paper also describes a deep learning-based quality control module that recovers additional peptides, including peptides mapped to the genome but absent from reference proteomes.

Significance. If the reported results are accurate, pUniFind would represent a substantial advance in computational proteomics: a single model that unifies database search rescoring and open de novo sequencing, with demonstrated gains across diverse instruments and applications. The scale of the training data, the breadth of validation strategies (entrapment, metabolic labeling, mixed-species searches, and genome-translated database matching), and the explicit design to avoid label leakage by decoupling representation learning from scoring are notable strengths. The claim of handling over 1,300 modifications in a zero-shot de novo setting is particularly significant and would address a long-standing limitation of existing de novo tools. However, the evaluation's reliance on Open-pFind as both the training annotator and the ground-truth labeler for de novo accuracy creates a substantial circularity risk that must be resolved before the central claims can be accepted.

major comments (4)
  1. [Fig. 1 and training description] The de novo accuracy evaluation is circular: the training set is annotated by Open-pFind, and the modification-level, site-level, and sequence-level accuracies in Fig. 5c–e are defined by agreement with Open-pFind identifications. The reported average peptide-level recall of 63.8% on the 21PTM dataset and the 60% improvement over pNovo are therefore measurements of agreement with the annotator, not independent evidence of correct de novo sequencing. The paper should include validation against an independent ground truth—for example, synthetic peptide spectra with known modifications, a cross-engine consensus that excludes Open-pFind, or a clearly documented exclusion of the 21PTM spectra from the training set. Without such a control, the central de novo claims remain unverified.
  2. [Results: 'Application of pUniFind in open de novo sequencing'] The text does not explicitly state that the 21PTM evaluation spectra were excluded from the 6,524 files used for training. If any of these spectra or their near-identical counterparts appeared in training, the reported recall and accuracy numbers would be inflated by memorization. This is a load-bearing issue for the de novo claims; the authors must either provide an explicit exclusion statement or quantify the overlap.
  3. [Results: 'Performance evaluation on MS/MS data from various species'] The performance gains across the nine-species datasets are reported as a single value per dataset without error bars, replicate runs, or statistical significance testing. The reported improvements range from 2% to 18% over Open-pFind, and the variability across datasets is substantial. Given that the database search workflow uses Open-pFind as the candidate generator and pUniFind as the rescoring model, the paper should demonstrate that the observed gains are reproducible and not driven by a few spectra or by the specific choice of the top-k prefilter (10 or 20). At minimum, a per-dataset breakdown of the number of spectra and the variance across subsets should be provided.
  4. [Results: 'The pUniFind model and its integration into the database search workflow'] The workflow rescores only spectra whose top-ranked candidate has a q-value below 0.1 from Open-pFind, and then applies target-decoy analysis to the final pUniFind scores. The validity of this two-step FDR control depends on the assumption that pUniFind's scores are well-calibrated and that the prefilter does not distort the target-decoy ratio. The paper should report the target-decoy score distributions for pUniFind (e.g., as shown for timsTOF in Fig. 3d) for the database search results, and should justify the q=0.1 prefilter threshold. Without this, the reported false discovery rates cannot be verified.
minor comments (5)
  1. [Introduction] The abstract states that pUniFind is 'the first large-scale multimodal pre-trained model in proteomics' and 'the first deep learning-based open de novo sequencing method.' The claims of novelty should be qualified by placing them in the context of recent work such as DeepSearch, DDA-BERT, and Casanovo, and by clarifying the specific sense in which 'unified' is used (the model itself performs both tasks, but the database search workflow still relies on Open-pFind for candidate generation).
  2. [Fig. 1] The architecture description in Fig. 1 mentions a 'Peptide Length Aware (PLA) module' but the main text does not define PLA or explain how length conditioning is implemented. Please provide a clear description in the Methods section.
  3. [Results] The sentence 'Tesorai slightly outperformed conventional search engines' is vague; please specify which engines and datasets are compared, and provide the corresponding numbers in the supplement.
  4. [Fig. 5] The text says 'the target PTM ranked within the top four by number of identified PSMs, and was the most frequently identified modification in 81% of the datasets,' but the corresponding ranking for the remaining 19% is not shown. Please include the complete ranking table in the supplement.
  5. [Throughout] Several references contain placeholder question marks (e.g., 'SEQUEST ?', 'Alphapept ?', 'Tesorai ?', '8'). Please resolve these. Also, 'a a 60% improvement' on page 10 contains a duplicated article. These should be corrected before publication.

Circularity Check

2 steps flagged · score 4.0 of 10

Open de novo recall and modification accuracy are scored against Open-pFind labels, the same engine that generated the training set; external anchors keep the core database-search claim independent.

  1. self definitional [Results, 'Application of pUniFind in open de novo sequencing of diverse modification-rich and regular datasets'; Fig. 5c–e caption]
    "Modification level accuracy is defined as follows: for all PSMs with modifications beyond Carbamidomethyl[C], a PSM is considered correct if pUniFind predicts the exact same modifications as those identified by Open-pFind. Similarly, site-level accuracy is assessed for all PSMs with modifications beyond Carbamidomethyl[C]. In this case, a PSM is deemed correct if both the modifications and their sites, as predicted by pUniFind, match exactly with those determined by Open-pFind."

    Open-pFind is the same engine used to annotate the 100M-PSM training set ('We employed Open-pFind to annotate a large-scale MS/MS dataset via open database search'), and the training objective uses Open-pFind's top-ranked candidate as positive and ranks 3–10 as negatives. Defining de novo 'accuracy' as exact agreement with Open-pFind therefore rewards learning the annotator's biases; any systematic Open-pFind error is scored as correct. The paper reports 63.8% average recall and a 60% improvement over pNovo on this basis. The text does not state that the 21PTM spectra were excluded from the 6,524 training files, so for these metrics the 'prediction' may be agreement with the training signal rather than independent de novo accuracy.

  2. other [Results, 'Application of pUniFind in open de novo sequencing of diverse modification-rich and regular datasets'; Fig. 5h sentence]
    "Additionally, the six most abundant modifications identified by pUniFind in the open de novo sequencing approach were identical to those reported by Open-pFind (Fig. 5h), thereby confirming the robustness of the modification discovery process."

    This validation is circular: Open-pFind supplied the training labels, so a match between pUniFind and Open-pFind on the most abundant modifications is expected if the model learned the annotator's modification distribution. It does not independently confirm the discovered modifications; independent confirmation would require synthetic peptides with known modifications or cross-validation against an engine not used in training.

full rationale

The paper's central database-search claims are substantially supported by independent checks: entrapment databases with species absent from training, metabolic-labeling NaN ratios, mixed-species timsTOF experiments, and genome/binding-affinity mapping do not rely on Open-pFind labels. These external anchors mean the core claim is not wholly circular. However, the open de novo evaluation—21PTM recall, modification-level and site-level accuracy, and modification-rank agreement—is defined against Open-pFind labels, the same tool that annotated the 100M-PSM training set. With no stated exclusion of the 21PTM spectra from training, those 'predictions' reduce in part to agreement with the training annotator rather than independent ground truth. The paper itself concedes that conventional database-search-based de novo evaluation is 'inherently limited,' yet uses exactly that framework for headline de novo numbers. Score 4 rather than higher because the database-search gains and several validation strategies are independent of the annotator loop.

Assumptions & free parameters 6 free parameters · 6 assumptions · 0 invented entities

The central claims rest on large empirical training data, indirect validation proxies, and hand-chosen workflow thresholds, rather than on formal derivation. No new physical or biological entities are introduced. The most important unpaid assumptions are the correctness of Open-pFind labels and the validity of target-decoy FDR control.

free parameters (6)
  • Top-k candidate prefilter size = 10 default; 20 for non-tryptic cases
    Chosen by hand in the database search workflow; restricts the candidate set that pUniFind can rescore.
  • q-value threshold for rescoring = 0.1
    Spectra whose top Open-pFind candidate has q-value above 0.1 are not rescored, which affects recall and precision.
  • De novo peptide length window = ±2 amino acids
    Peptide length is predicted first and sequences are generated only within 2 residues of that value, so length errors beyond 2 cannot be corrected.
  • Number of modifications used in modification-enriched de novo rescoring = top four predicted modifications plus user-specified ones
    pFind rescoring uses only four automatic modification types; performance depends on this selection.
  • Quality-control filtering thresholds = not disclosed; default filtering standard with no missing fragment ion sites
    The deep learning QC filter uses end-to-end scores, predicted spectra, and retention times, but exact decision thresholds are not given.
  • Model weights and training hyperparameters = not disclosed
    All performance claims depend on the fitted network weights and hyperparameters, which are not provided in the text.
assumptions (6)
  • domain assumption Open-pFind annotations are reliable ground truth for training and for most recall evaluations.
    Training set and de novo recall labels come from Open-pFind, a tool from the same lab; if labels contain systematic errors, pUniFind inherits them.
  • domain assumption Target-decoy FDR estimation remains valid for pUniFind scores.
    Final identifications use target-decoy analysis, which assumes the decoy distribution is a valid null and the model has no label leakage.
  • domain assumption The learned spectrum and peptide embeddings capture real fragmentation biochemistry.
    The pre-training tasks, such as spectrum prediction, fragment ion type, and peptide length, are assumed to teach physically meaningful features rather than dataset artifacts.
  • domain assumption Indirect validation proxies are sufficient evidence of identification accuracy.
    Entrapment ratios, metabolic labeling NaN ratios, genomic mapping, and HLA binding are indirect; each has caveats such as database composition and incomplete HLA binding predictions.
  • domain assumption Fine-tuning on timsTOF and Astral data with Open-pFind-derived labels does not bias the validation results.
    The one-epoch fine-tuning collects positives and negatives from Open-pFind, so timsTOF and Astral gains could partly reflect the same annotator's preferences.
  • domain assumption Cross-instrument generalization from DDA training data to Astral and timsTOF is valid.
    The default model is retrained with timsTOF data but applied to Astral; instrument-specific effects such as ion mobility are not explicitly modeled.

how reviews work

0 comments
Cite this review

Pith. "Pith review of pUniFind: a unified large pre-trained deep learning model pushing the limit of mass spectra interpretation." pith.science (2026). https://pith.science/paper/73PXFCYL

@misc{pith2026250700087,
  author       = {Pith},
  title        = {Pith review of: pUniFind: a unified large pre-trained deep learning model pushing the limit of mass spectra interpretation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/73PXFCYL}},
  note         = {Machine review of arXiv:2507.00087}
}
read the original abstract

Deep learning has advanced mass spectrometry data interpretation, yet most models remain feature extractors rather than unified scoring frameworks. We present pUniFind, the first large-scale multimodal pre-trained model in proteomics that integrates end-to-end peptide-spectrum scoring with open, zero-shot de novo sequencing. Trained on over 100 million open search-derived spectra, pUniFind aligns spectral and peptide modalities via cross modality prediction and outperforms traditional engines across diverse datasets, particularly achieving a 42.6 percent increase in the number of identified peptides in immunopeptidomics. Supporting over 1,300 modifications, pUniFind identifies 60 percent more PSMs than existing de novo methods despite a 300-fold larger search space. A deep learning based quality control module further recovers 38.5 percent additional peptides including 1,891 mapped to the genome but absent from reference proteomes while preserving full fragment ion coverage. These results establish a unified, scalable deep learning framework for proteomic analysis, offering improved sensitivity, modification coverage, and interpretability.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

27 extracted references · 26 canonical work pages

  1. [1]

    Nature 537(7620), 347–355 (2016)

    Aebersold, R., Mann, M.: Mass-spectrometric exploration of proteome structure and function. Nature 537(7620), 347–355 (2016)

    2016

  2. [2]

    Nature biotechnology 36(11), 1059–1061 (2018)

    Chi, H., Liu, C., Yang, H., Zeng, W.-F., Wu, L., Zhou, W.-J., Wang, R.-M., Niu, X.-N., Ding, Y .-H., Zhang, Y .,et al.: Comprehensive identification of peptides in tandem mass spectra using an efficient open search engine. Nature biotechnology 36(11), 1059–1061 (2018)

    work page 2018

  3. [3]

    Nature biotechnology 26(12), 1367–1372 (2008)

    Cox, J., Mann, M.: Maxquant enables high peptide identification rates, individualized ppb-range mass accuracies and proteome-wide protein quantification. Nature biotechnology 26(12), 1367–1372 (2008)

    work page 2008

  4. [4]

    Nature methods 14(5), 513–520 (2017)

    Kong, A.T., Leprevost, F.V ., Avtonomov, D.M., Mellacheruvu, D., Nesvizhskii, A.I.: Msfragger: ultrafast and comprehensive peptide identification in mass spectrometry–based proteomics. Nature methods 14(5), 513–520 (2017)

    work page 2017

  5. [5]

    Journal of proteome research16(2), 645–654 (2017)

    Yang, H., Chi, H., Zhou, W.-J., Zeng, W.-F., He, K., Liu, C., Sun, R.-X., He, S.-M.: Open-pnovo: de novo peptide sequencing with thousands of protein modifications. Journal of proteome research16(2), 645–654 (2017)

    work page 2017

  6. [6]

    Bioinformatics 35(14), 183–190 (2019)

    Yang, H., Chi, H., Zeng, W.-F., Zhou, W.-J., He, S.-M.: pnovo 3: precise de novo peptide sequencing using a learning-to-rank framework. Bioinformatics 35(14), 183–190 (2019)

    work page 2019

  7. [7]

    Rapid communications in mass spectrometry 17(20), 2337–2342 (2003)

    Ma, B., Zhang, K., Hendrie, C., Liang, C., Li, M., Doherty-Kirby, A., Lajoie, G.: Peaks: powerful software for peptide de novo sequencing by tandem mass spectrometry. Rapid communications in mass spectrometry 17(20), 2337–2342 (2003)

    work page 2003

  8. [8]

    Analytical chemistry 77(4), 964–973 (2005)

    Frank, A., Pevzner, P.: Pepnovo: de novo peptide sequencing via probabilistic network modeling. Analytical chemistry 77(4), 964–973 (2005)

    work page 2005

Show all 27 references
  1. [9]

    Proceedings of the National Academy of Sciences 114(31), 8247–8252 (2017)

    Tran, N.H., Zhang, X., Xin, L., Shan, B., Li, M.: De novo peptide sequencing by deep learning. Proceedings of the National Academy of Sciences 114(31), 8247–8252 (2017)

    2017

  2. [10]

    Nature Machine Intelligence 3(5), 420–425 (2021)

    Qiao, R., Tran, N.H., Xin, L., Chen, X., Li, M., Shan, B., Ghodsi, A.: Computationally instrument- resolution-independent de novo peptide sequencing for high-resolution devices. Nature Machine Intelligence 3(5), 420–425 (2021)

    2021

  3. [11]

    In: International Conference on Machine Learning, pp

    Yilmaz, M., Fondrie, W., Bittremieux, W., Oh, S., Noble, W.S.: De novo mass spectrometry peptide sequencing with a transformer model. In: International Conference on Machine Learning, pp. 25514– 25522 (2022). PMLR

    2022

  4. [12]

    Nature communications 15(1), 6427 (2024)

    Yilmaz, M., Fondrie, W.E., Bittremieux, W., Melendez, C.F., Nelson, R., Ananth, V ., Oh, S., Noble, W.S.: Sequence-to-sequence translation from mass spectra to peptides with a transformer model. Nature communications 15(1), 6427 (2024)

    2024

  5. [13]

    Nature Communications 13(1), 7238 (2022)

    Zeng, W.-F., Zhou, X.-X., Willems, S., Ammar, C., Wahle, M., Bludau, I., V oytik, E., Strauss, M.T., Mann, M.: Alphapeptdeep: a modular deep learning framework to predict peptide properties for proteomics. Nature Communications 13(1), 7238 (2022)

    2022

  6. [14]

    Nature Communications 14(1), 4539 (2023)

    Yang, K.L., Yu, F., Teo, G.C., Li, K., Demichev, V ., Ralser, M., Nesvizhskii, A.I.: Msbooster: improv- ing peptide identification rates using deep learning-based features. Nature Communications 14(1), 4539 (2023)

    2023

  7. [15]

    Nature methods 16(6), 509–518 (2019) 17

    Gessulat, S., Schmidt, T., Zolg, D.P., Samaras, P., Schnatbaum, K., Zerweck, J., Knaute, T., Rechen- berger, J., Delanghe, B., Huhmer, A., et al.: Prosit: proteome-wide prediction of peptide tandem mass spectra by deep learning. Nature methods 16(6), 509–518 (2019) 17

    2019

  8. [16]

    Journal of Proteomics 251, 104414 (2022)

    Zhou, W.-J., Wei, Z.-H., He, S.-M., Chi, H.: pvalid 2: A deep learning based validation method for pep- tide identification in shotgun proteomics with increased discriminating power. Journal of Proteomics 251, 104414 (2022)

    2022

  9. [17]

    Molecular & Cellular Proteomics 17(9), 1850–1863 (2018)

    Zolg, D.P., Wilhelm, M., Schmidt, T., Médard, G., Zerweck, J., Knaute, T., Wenschuh, H., Reimer, U., Schnatbaum, K., Kuster, B.: Proteometools: Systematic characterization of 21 post-translational protein modifications by liquid chromatography tandem mass spectrometry (lc-ms/m...

    2018

  10. [18]

    In: International Conference on Machine Learning, pp

    Li, J., Li, D., Xiong, C., Hoi, S.: Blip: Bootstrapping language-image pre-training for unified vision- language understanding and generation. In: International Conference on Machine Learning, pp. 12888–12900 (2022). PMLR

    2022

  11. [19]

    IEEE Photonics Journal 13(4), 1–6 (2021)

    Zhao, J., Zhang, Z., Li, Y ., Cen, L.: Multidimensional information assisted deep learning realizing flexible recognition of vortex beam modes. IEEE Photonics Journal 13(4), 1–6 (2021)

    2021

  12. [20]

    bioRxiv, 2024–07 (2024)

    Hayes, T., Rao, R., Akin, H., Sofroniew, N.J., Oktay, D., Lin, Z., Verkuil, R., Tran, V .Q., Deaton, J., Wiggert, M., et al.: Simulating 500 million years of evolution with a language model. bioRxiv, 2024–07 (2024)

    2024

  13. [21]

    In: Forty-first International Conference on Machine Learning

    Zhuang, W., Song, J., Li, Y ., Lu, S., et al.: Pre-training protein bi-level representation through span mask strategy on 3d protein chains. In: Forty-first International Conference on Machine Learning

  14. [22]

    Zhou, G., Gao, Z., Ding, Q., Zheng, H., Xu, H., Wei, Z., Zhang, L., Ke, G.: Uni-mol: A universal 3d molecular representation learning framework (2023)

    2023

  15. [23]

    Molecular & Cellular Proteomics 20 (2021)

    Willems, S., V oytik, E., Skowronek, P., Strauss, M.T., Mann, M.: Alphatims: indexing trapped ion mobility spectrometry–tof data for fast and easy accession and visualization. Molecular & Cellular Proteomics 20 (2021)

    2021

  16. [24]

    Nature624(7990), 192–200 (2023)

    Michaelis, A.C., Brunner, A.-D., Zwiebel, M., Meier, F., Strauss, M.T., Bludau, I., Mann, M.: The social and structural architecture of the yeast protein interactome. Nature624(7990), 192–200 (2023)

    2023

  17. [25]

    Journal of Proteome Research 23(10), 4761–4769 (2024)

    Melendez, C., Sanders, J., Yilmaz, M., Bittremieux, W., Fondrie, W.E., Oh, S., Noble, W.S.: Account- ing for digestion enzyme bias in casanovo. Journal of Proteome Research 23(10), 4761–4769 (2024)

    2024

  18. [26]

    Journal of the American Society for Mass Spectrometry 31(7), 1473–1482 (2020)

    Cheng, K., Ning, Z., Zhang, X., Li, L., Liao, B., Mayne, J., Figeys, D.: Metalab 2.0 enables accu- rate post-translational modifications profiling in metaproteomics. Journal of the American Society for Mass Spectrometry 31(7), 1473–1482 (2020)

    2020

  19. [27]

    In: Proceedings of the IEEE/CVF International Conference on Computer Vision, pp

    Ridnik, T., Ben-Baruch, E., Zamir, N., Noy, A., Friedman, I., Protter, M., Zelnik-Manor, L.: Asym- metric loss for multi-label classification. In: Proceedings of the IEEE/CVF International Conference on Computer Vision, pp. 82–91 (2021) 18 Data availability All raw data used f...

    2021

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.