Introducing<i>π</i>-HelixNovo for practical large-scale de novo peptide sequencing

Yang, Tingpeng; Ling, Tianze; Sun, Baobao; Liang, Zhendong; Xu, Fan; Huang, Xiansong; Xie, Likun; He, Yonghong; Li, Leyuan; He, Fuchu; Yu, Wang; Chang, Cheng

doi:10.1101/2023.07.15.549133

Cited by 3 publications

(2 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The potential for deep learning methods to improve our ability to perform de novo sequencing has now been widely recognized. While this paper was under review, at least six additional deep learning de novo sequencing methods have been published, including GraphNovo [38], PepNet [39], Denovo-GCN [40], Spectralis [41], π-HelixNovo [42], and NovoB [43]. Clearly, the field would benefit from an exhaustive and rigorous benchmark comparison of this growing field of tools.…”

Section: Discussionmentioning

confidence: 99%

Sequence-to-sequence translation from mass spectra to peptides with a transformer model

Fondrie

Yilmaz

Bittremieux

et al. 2023

Preprint

View full text Add to dashboard Cite

A fundamental challenge for any mass spectrometry-based proteomics experiment is the identification of the peptide that generated each acquired tandem mass spectrum. Although approaches that leverage known peptide sequence databases are widely used and effective for well-characterized model organisms, such methods cannot detect unexpected peptides and can be impractical or impossible to apply in some settings. Thus, the ability to assign peptide sequences to the acquired tandem mass spectra without prior information -- de novo peptide sequencing -- is valuable for gaining biological insights for tasks including antibody sequencing, immunopeptidomics, and metaproteomics. Although many methods have been developed to address this de novo sequencing problem, it remains an outstanding challenge, in part due to the difficulty of modeling the irregular data structure of tandem mass spectra. Here, we describe Casanovo, a machine learning model that uses a transformer neural network architecture to translate the sequence of peaks in a tandem mass spectrum into the sequence of amino acids that comprise the generating peptide. We train a Casanovo model from 30 million labeled spectra and demonstrate that the model outperforms several state-of-the-art methods on a cross-species benchmark dataset. We also develop a version of Casanovo that is fine-tuned for non-enzymatic peptides. Finally, we demonstrate that Casanovo's superior performance improves the analysis of immunopeptideomics and metaproteomics experiments and allows us to delve deeper into the dark proteome.

show abstract

Section: Discussionmentioning

confidence: 99%

Sequence-to-sequence translation from mass spectra to peptides with a transformer model

Fondrie

Yilmaz

Bittremieux

et al. 2023

Preprint

View full text Add to dashboard Cite

show abstract

“…However, the possibility of sequencing mistakes cannot be disregarded and in practice, external databases and complementary DBS are commonly used. Recently, new dnS programs like Casanovo (Yilmaz et al, 2022), InstaNovo (Eloff et al, 2023), or π-HelixNovo (Yang et al, 2023) are reporting promising performance. They use transformer neural network architecture, (as used in Large Language Models), which seem suited to protein sequences as they can be seen as another language.…”

Section: Discussionmentioning

confidence: 99%

Benchmarking the identification of a single degraded protein to explore optimal search strategies for ancient proteins

Rodriguez Palomo,

Nair,

Chiang

et al. 2023

Preprint

View full text Add to dashboard Cite

Palaeoproteomics is a rapidly evolving discipline, and practitioners are constantly developing novel strategies for the analyses and interpretations of complex, degraded protein mixtures. The community has also established standards of good practice to interrogate our data. However, there is a lack of a systematic exploration of how these affect the identification of peptides, post-translational modifications (PTMs) and protein and their significance (through the False Discovery Rate) and correctness. We systematically investigated the performance of a wide range of sequencing tools and search engines in a controlled system: the experimental degradation of the single purified bovine β-lactoglobulin (BLG), heated at 95 °C and pH 7 for 0, 4 and 128 days. We target BLG since it is one of the most robust and ubiquitous proteins in the archaeological record. We tested different reference database choices, a targeted dairy protein one, and whole bovine proteome and three digestion options (tryptic-, semi-tryptic– and non-specific searches), in order to evaluate the effects of search space and the identification of peptides. We also explored alternative strategies, including open search that allows for the global identification of PTMs based upon wide precursor mass tolerance andde novosequencing to boost sequence coverage. We analysed the samples using Mascot, MaxQuant, Metamorpheus, pFind, Fragpipe and DeNovoGUI (pepNovo, DirecTag, Novor), and we benchmarked these tools and discussed the optimal strategy for the characterisation of ancient proteins.

show abstract

π-PrimeNovo: An Accurate and Efficient Non-Autoregressive Deep Learning Model for De Novo Peptide Sequencing

Zhang,

Ling,

Jin

et al. 2024

Preprint

Self Cite

View full text Add to dashboard Cite

Peptide sequencing via tandem mass spectrometry (MS/MS) is fundamental in proteomics data analysis, playing a pivotal role in unraveling the complex world of proteins within biological systems. In contrast to conventional database searching methods, deep learning models excel in de novo sequencing peptides absent from existing databases, thereby facilitating the identification and analysis of novel peptide sequences. Current deep learning models for peptide sequencing predominantly use an autoregressive generation approach, where early errors can cascade, largely affecting overall sequence accuracy. And the usage of sequential decoding algorithms such as beam search suffers from the low inference speed. To address this, we introduce π-PrimeNovo, a non-autoregressive Transformer-based deep learning model designed to perform accurate and efficient de novo peptide sequencing. With the proposed novel architecture, π-PrimeNovo achieves significantly higher accuracy and up to 69x faster sequencing compared to the state-of-the-art methods. This remarkable speed makes it highly suitable for computation-extensive peptide sequencing tasks such as metaproteomic research, where π-PrimeNovo efficiently identifies the microbial species-specific peptides. Moreover, π-PrimeNovo has been demonstrated to have a powerful capability in accurately mining phosphopeptides in a non-enriched phosphoproteomic dataset, showing an alternative solution to detect low-abundance post-translational modifications (PTMs). We suggest that this work not only advances the development of peptide sequencing techniques but also introduces a transformative computational model with wide-range implications for biological research.

show abstract

Introducingπ-HelixNovo for practical large-scale de novo peptide sequencing

Cited by 3 publications

References 17 publications

Sequence-to-sequence translation from mass spectra to peptides with a transformer model

Sequence-to-sequence translation from mass spectra to peptides with a transformer model

Benchmarking the identification of a single degraded protein to explore optimal search strategies for ancient proteins

π-PrimeNovo: An Accurate and Efficient Non-Autoregressive Deep Learning Model for De Novo Peptide Sequencing

Contact Info

Product

Resources

About