Toward an Integrated Machine Learning Model of a Proteomics Experiment

Neely, Benjamin A.; Dorfer, Viktoria; Martens, Lennart; Bludau, Isabell; Bouwmeester, Robbin; Degroeve, Sven; Deutsch, Eric W.; Gessulat, Siegfried; Käll, Lukas; Palczynski, Pawel; Payne, Samuel H.; Rehfeldt, Tobias Greisager; Schmidt, Tobias; Schwämmle, Veit; Uszkoreit, Julian; Vizcaíno, Juan Antonio; Wilhelm, Mathias; Palmblad, Magnus

doi:10.1021/acs.jproteome.2c00711

Cited by 37 publications

(20 citation statements)

References 141 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Our experiments showcase the stability and effectiveness of PepT3 in improving the current paradigm for deep MS/MS spectrum prediction. As benchmark data sets are being established and more DSPMs are emerging, PepT3 can serve as a critical link in connecting DSPMs from standardized training to laboratory evaluation. DSPM with PepT3 holds great potential for proteomic research, particularly for highly complex and quantity-limited samples.…”

Section: Discussionmentioning

confidence: 99%

“…Therefore, supervised learning-based training paradigm used by these models has a limitation. 17 Variations in the characteristics of mass spectra from laboratory to laboratory and experiment to experiment 18 can lead to a generalization problem for the out-of-distribution experimental spectra. Empirically, this could weaken DSPM's effectiveness in predicting spectra and result in a degradation of performance.…”

Section: ■ Introductionmentioning

confidence: 99%

“…and more DSPMs are emerging,17 PepT3 can serve as a critical link in connecting DSPMs from standardized training to laboratory evaluation. DSPM with PepT3 holds great potential for proteomic research, particularly for highly complex and quantity-limited samples.…”

mentioning

confidence: 99%

See 2 more Smart Citations

Test-Time Training for Deep MS/MS Spectrum Prediction Improves Peptide Identification

Ye,

He,

Wang

et al. 2023

J. Proteome Res.

View full text Add to dashboard Cite

In bottom-up proteomics, peptide-spectrum matching is critical for peptide and protein identification. Recently, deep learning models have been used to predict tandem mass spectra of peptides, enabling the calculation of similarity scores between the predicted and experimental spectra for peptide-spectrum matching. These models follow the supervised learning paradigm, which trains a general model using paired peptides and spectra from standard data sets and directly employs the model on experimental data. However, this approach can lead to inaccurate predictions due to differences between the training data and the experimental data, such as sample types, enzyme specificity, and instrument calibration. To tackle this problem, we developed a test-time training paradigm that adapts the pretrained model to generate experimental data-specific models, namely, PepT3. PepT3 yields a 10−40% increase in peptide identification depending on the variability in training and experimental data. Intriguingly, when applied to a patient-derived immunopeptidomic sample, PepT3 increases the identification of tumor-specific immunopeptide candidates by 60%. Two-thirds of the newly identified candidates are predicted to bind to the patient's human leukocyte antigen isoforms. To facilitate access of the model and all the results, we have archived all the intermediate files in Zenodo.org with identifier 8231084.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: ■ Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Test-Time Training for Deep MS/MS Spectrum Prediction Improves Peptide Identification

Ye,

He,

Wang

et al. 2023

J. Proteome Res.

View full text Add to dashboard Cite

show abstract

“…The filtering was performed using the filter API using the -q option with MongoDB or string filters available in the GitHub repository. The model training and testing were performed using the network API with "-t prosit -e 100 -sos -s n" to train a Prosit model for 100 We utilized a Bayesian approximation of the model uncertainty, by performing model inference with dropout enabled [25,26]. The real retention time values are then plotted against the mean predicted values, with the color of the data point corresponding to the normalized variances of the predicted values.…”

Section: Methodsmentioning

confidence: 99%

Variance Analysis of LC-MS Experimental Factors and Their Impact on Machine Learning

Rehfeldt

Krawczyk

Gregersen

et al. 2023

Preprint

Self Cite

View full text Add to dashboard Cite

Background Machine learning (ML) technologies, especially deep learning (DL), have gained increasing attention in predictive mass spectrometry (MS) for enhancing the data processing pipeline from raw data analysis to end-user predictions and re-scoring. ML models need large-scale datasets for training and re-purposing, which can be obtained from a range of public data repositories. However, applying ML to public MS datasets on larger scales is challenging, as they vary widely in terms of data acquisition methods, biological systems, and experimental designs. Results We aim to facilitate ML efforts in MS data by conducting a systematic analysis of the potential sources of variance in public MS repositories. We also examine how these factors affect ML performance and perform a comprehensive transfer learning to evaluate the benefits of current best practice methods in the field for transfer learning. Conclusions Our findings show significantly higher levels of homogeneity within a project than between projects, which indicates that it's important to construct datasets most closely resembling future test cases, as transferability is severely limited for unseen datasets. We also found that transfer learning, although it did increase model performance, did not increase model performance compared to a non-pre-trained model.

show abstract

“…Initially popularized in fields like medical imaging, speech recognition, computer vision, and natural language processing, these algorithms have marked milestones such as predicting folding of proteins with remarkable accuracy, making them particularly effective when applied to large and complex data 25 . Given the data-intensive nature of modern biotechnological research, proteomics is increasingly becoming a fertile ground for the application of deep learning technologies [26][27][28] .…”

Section: Introductionmentioning

confidence: 99%

Deep Learning for Protein–peptide binding Prediction: Incorporating Sequence, Structural and Language Model Features

Chandra,

Sharma,

Dehzangi

et al. 2023

Preprint

View full text Add to dashboard Cite

Protein-peptide interactions play a crucial role in various cellular processes and are implicated in abnormal cellular behaviors leading to diseases such as cancer. Therefore, understanding these interactions is vital for both functional genomics and drug discovery efforts. Despite a significant increase in the availability of protein-peptide complexes, experimental methods for studying these interactions remain laborious, time-consuming, and expensive. Computational methods offer a complementary approach but often fall short in terms of prediction accuracy. To address these challenges, we introduce PepCNN, a deep learning-based prediction model that incorporates structural and sequence-based information from primary protein sequences. By utilizing a combination of half-sphere exposure, position specific scoring matrices, and pre-trained transformer language model, PepCNN outperforms state-of-the-art methods in terms of specificity, precision, and AUC. The PepCNN software and datasets are publicly available at https://github.com/abelavit/PepCNN.git.

show abstract

Toward an Integrated Machine Learning Model of a Proteomics Experiment

Cited by 37 publications

References 141 publications

Test-Time Training for Deep MS/MS Spectrum Prediction Improves Peptide Identification

Test-Time Training for Deep MS/MS Spectrum Prediction Improves Peptide Identification

Variance Analysis of LC-MS Experimental Factors and Their Impact on Machine Learning

Deep Learning for Protein–peptide binding Prediction: Incorporating Sequence, Structural and Language Model Features

Contact Info

Product

Resources

About