Accurate De Novo Prediction of Protein Contact Map by Ultra-Deep Learning Model

Wang, Sheng; Sun, Siqi; Li, Zhen; Zhang, Renyu; Xu, Jinbo

doi:10.1101/073239

Cited by 151 publications

(276 citation statements)

References 45 publications

Supporting

Mentioning

273

Contrasting

Order By: Relevance

“…It now forms the core of the artificial intelligence platforms of several technology giants, such as Google, Facebook, and Microsoft, as well as many startups in the industry. Deep learning has also made its way into biological sciences (30) [for instance, in the field of genomics, where deep neural network models have been developed for predicting the effects of noncoding single-nucleotide variants (31), predicting protein DNA and RNA binding sites (32), protein contact map prediction (33), and MS imaging (34)]. The key aspect of deep learning is its ability to learn multiple levels of representation of high-dimensional data through its many layers of neurons.…”

mentioning

confidence: 99%

De novo peptide sequencing by deep learning

Tran

Zhang

Lei

et al. 2017

Proc. Natl. Acad. Sci. U.S.A.

352

375

View full text Add to dashboard Cite

De novo peptide sequencing from tandem MS data is the key technology in proteomics for the characterization of proteins, especially for new sequences, such as mAbs. In this study, we propose a deep neural network model, DeepNovo, for de novo peptide sequencing. DeepNovo architecture combines recent advances in convolutional neural networks and recurrent neural networks to learn features of tandem mass spectra, fragment ions, and sequence patterns of peptides. The networks are further integrated with local dynamic programming to solve the complex optimization task of de novo sequencing. We evaluated the method on a wide variety of species and found that DeepNovo considerably outperformed state of the art methods, achieving 7.7-22.9% higher accuracy at the amino acid level and 38.1-64.0% higher accuracy at the peptide level. We further used DeepNovo to automatically reconstruct the complete sequences of antibody light and heavy chains of mouse, achieving 97.5-100% coverage and 97.2-99.5% accuracy, without assisting databases. Moreover, DeepNovo is retrainable to adapt to any sources of data and provides a complete end-to-end training and prediction solution to the de novo sequencing problem. Not only does our study extend the deep learning revolution to a new field, but it also shows an innovative approach in solving optimization problems by using deep learning and dynamic programming.deep learning | MS | de novo sequencing P roteomics research focuses on large-scale studies to characterize the proteome, the entire set of proteins, in a living organism (1-5). In proteomics, de novo peptide sequencing from tandem MS data plays the key role in the characterization of novel protein sequences. This field has been actively studied over the past 20 y, and many de novo sequencing tools have been proposed, such as PepNovo, PEAKS, NovoHMM, MSNovo, pNovo, UniNovo, and Novor among others (6-19). The recent "gold rush" into mAbs has undoubtedly elevated the application of de novo sequencing to a new horizon (20-23). However, computational challenges still remain, because MS/MS spectra contain much noise and ambiguity that require rigorous global optimization with various forms of dynamic programming that have been developed over the past decade (8-10, 12, 13, 15-19, 24).In this study, we introduce neural networks and deep learning to de novo peptide sequencing and achieve major breakthroughs on this well-studied problem. Deep learning has recently brought about a revolution in many research fields (25), repeatedly breaking state of the art records in image processing (26, 27), speech recognition (28), and natural language processing (29). It now forms the core of the artificial intelligence platforms of several technology giants, such as Google, Facebook, and Microsoft, as well as many startups in the industry. Deep learning has also made its way into biological sciences (30) [for instance, in the field of genomics, where deep neural network models have been developed for predicting the effects of noncoding single-nucleotide ...

show abstract

mentioning

confidence: 99%

De novo peptide sequencing by deep learning

Tran

Zhang

Lei

et al. 2017

Proc. Natl. Acad. Sci. U.S.A.

352

375

View full text Add to dashboard Cite

show abstract

“…Deep learning has recently been applied successfully to several important biological problems (Kelley et al, 2016;Zhou and Troyanskaya, 2015;Wang et al, 2017;Leung et al, 2014). Such methods are attractive due to their ability to extract complex yet meaningful feature representations from massive data sets without requiring extensive feature engineering.…”

Section: Introductionmentioning

confidence: 99%

Nucleotide sequence and DNaseI sensitivity are predictive of 3D chromatin architecture

Schreiber

Libbrecht

Bilmes

et al. 2017

Preprint

View full text Add to dashboard Cite

Motivation: Recently, Hi-C has been used to probe the 3D chromatin architecture of multiple organisms and cell types. The resulting collections of pairwise contacts across the genome have connected chromatin architecture to many cellular phenomena, including replication timing and gene regulation. However, high resolution (10 kb or finer) contact maps remain scarce due to the expense and time required for collection. A computational method for predicting pairwise contacts without the need to run a Hi-C experiment would be invaluable in understanding the role that 3D chromatin architecture plays in genome biology. Results: We describe Rambutan, a deep convolutional neural network that predicts Hi-C contacts at 1 kb resolution using nucleotide sequence and DNaseI assay signal as inputs. Specifically, Rambutan identifies locus pairs that engage in high confidence contacts according to Fit-Hi-C, a previously described method for assigning statistical confidence estimates to Hi-C contacts. We first demonstrate Rambutan's performance across chromosomes at 1 kb resolution in the GM12878 cell line. Subsequently, we measure Rambutan's performance across six cell types. In this setting, the model achieves an area under the receiver operating characteristic curve between 0.7662 and 0.8246 and an area under the precision-recall curve between 0.3737 and 0.9008. We further demonstrate that the predicted contacts exhibit expected trends relative to histone modification ChIP-seq data, replication timing measurements, and annotations of functional elements such as promoters and enhancers. Finally, we predict Hi-C contacts for 53 human cell types and show that the predictions cluster by cellular function. Availability: Tutorials and source code for Rambutan are publicly available at https://github.com/ jmschrei/rambutan.

show abstract

“…Map (CoRe-CMap) model, a meta-method for predicting protein motion from sequence. The CoRe-CMap approach combines residuewise predictions of protein disorder likelihoood with likelihoods of inter-residue contacts obtained from contact map-based protein structure prediction methods such as RaptorX 22 and CMAPpro. 23 By using likelihoods of disorder to weight a matrix of likelihoods of contact, CoRe-CMap tabulates likelihoods of "conformationally restraining" contacts.…”

Section: This Paper Presents the Conformationally Restraining Contactsmentioning

confidence: 99%

Prediction of protein flexibility using a conformationally restrained contact map

et al. 2018

View full text Add to dashboard Cite

Knowledge of protein flexibility is crucial to understanding protein function. However, probing protein flexibility by either experiment or computational simulations is a difficult process. In particular, many computational approaches to understanding protein flexibility require an experimentally determined protein structure. The Conformationally Restrained Contact Map (CoRe-CMap) approach reported here couples protein disorder predictions with protein structure predictions and only requires sequence data to predict protein flexibility. This paper reports the application of the CoRe-CMap model to predicting Lipari-Szabo order parameters of all proteins for which experimentally derived Lipari-Szabo order parameters are available in the BioMagResBank: the median root mean square deviation between a protein's predicted and experimentally derived order parameters is 0.124. Additionally, application of the CoRe-CMap model to predict Lipari-Szabo order parameters for the 10th Type III Domain in Fibronectin and a homologous domain from Tenascin demonstrates the ability of CoRe-CMap to predict functionally important differences in protein flexibility.

show abstract

Accurate De Novo Prediction of Protein Contact Map by Ultra-Deep Learning Model

Abstract: Motivation

Cited by 151 publications

References 45 publications

De novo peptide sequencing by deep learning

De novo peptide sequencing by deep learning

Nucleotide sequence and DNaseI sensitivity are predictive of 3D chromatin architecture

Prediction of protein flexibility using a conformationally restrained contact map

Contact Info

Product

Resources

About