DeepCNF-D: Predicting Protein Order/Disorder Regions by Weighted Deep Convolutional Neural Fields

Wang, Sheng; Weng, Shunyan; Ma, Jianzhu; Tang, Qingming

doi:10.3390/ijms160817315

Cited by 68 publications

(48 citation statements)

References 43 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Indeed, these neuron network architectures are dedicated to encode data of sequential nature. Indeed, they have been successfully designed and trained on proteomic data to achieve state-of-the-art performance in various prediction problems [62,63,64,65,66,67].…”

Section: Protein Sequence Encodermentioning

confidence: 99%

Evaluation of network architecture and data augmentation methods for deep learning in chemogenomics

Playe

Stoven

2019

Preprint

View full text Add to dashboard Cite

Among virtual screening methods that have been developed to facilitate the drug discovery process, chemogenomics presents the particularity to tackle the question of predicting ligands for proteins, at at scales both in the protein and chemical spaces. Therefore, in addition to to predict drug candidates for a given therapeutic protein target, like more classical ligand-based or receptor-based methods do, chemogenomics can also predict off-targets at the proteome level, and therefore, identify potential side-effects or drug repositioning opportunities. In this study, we study and compare machinelearning and deep learning approaches for chemogenomics, that are applicable to screen large sets of compounds against large sets of druggable proteins. State-of-the-art drug chemogenomics methods rely on expert-based chemical and protein descriptors or similarity measures. The recent development of deep learning approaches enabled to design algorithms that learn numerical abstract representations of molecular graphs and protein sequences in an end-to-end fashion, i.e., so that the learnt features optimise the objective function of the drug-target interaction prediction task. In this paper, we address drug-target interaction prediction at the druggable proteome-level, with what we define as the chemogenomic neuron network. This network consists of a feed-forward neuron network taking as input the combination of molecular and protein representations learnt by molecular graph and protein sequence encoders. We first propose a standard formulation of this chemogenomic neuron network. Then, we compare the performances of the standard chemogenomic network to reference deep learning or shallow (machine-learning without deep learning) methods. In particular, we show that such a representation learning approach is competitive with state-of-the-art chemogenomics with shallow methods, but not ultimately superior. We evaluate the most promising neuron network architectures and data augmentation techniques, such as multi-view and transfer learning, to improve the prediction performance of the chemogenomic network. Our results shed new insights on the design of chemogenomics approaches based on representation learning algorithms. Most importantly, we conclude from our observations that a promising research direction is to integrate heterogeneous sources of data such as various bioactivity datasets, or independently, multiple molecule and protein attribute views, instead of focusing on sophisticated, yet intuitively relevant, encoder's neuron network architecture.

show abstract

Section: Protein Sequence Encodermentioning

confidence: 99%

Evaluation of network architecture and data augmentation methods for deep learning in chemogenomics

Playe

Stoven

2019

Preprint

View full text Add to dashboard Cite

show abstract

“…However, as we will show in this paper, the accuracy of MWUTest is still very low, due to its limited capability to model transitions and to handle noise. In this work, we propose a novel algorithmic framework, DeepBound, to identify boundaries of expressed transcripts from reads alignment by using deep convolutional neural fields (DeepCNF) (Wang et al, 2015(Wang et al, , 2016c. Different from previous statistical methods, DeepBound can integrate a variety of information and automatically determine their quantity under different circumstances.…”

Section: Introductionmentioning

confidence: 99%

DeepBound: Accurate Identification of Transcript Boundaries via Deep Convolutional Neural Fields

Shao

Wang

2017

Preprint

Self Cite

View full text Add to dashboard Cite

Motivation: Reconstructing the full-length expressed transcripts (a.k.a. the transcript assembly problem) from the short sequencing reads produced by RNA-seq protocol plays a central role in identifying novel genes and transcripts as well as in studying gene expressions and gene functions. A crucial step in transcript assembly is to accurately determine the splicing junctions and boundaries of the expressed transcripts from the reads alignment. In contrast to the splicing junctions that can be efficiently detected from spliced reads, the problem of identifying boundaries remains open and challenging, due to the fact that the signal related to boundaries is noisy and weak. Results: We present DeepBound, an effective approach to identify boundaries of expressed transcripts from RNA-seq reads alignment. In its core DeepBound employs deep convolutional neural fields to learn the hidden distributions and patterns of boundaries. To accurately model the transition probabilities and to solve the label-imbalance problem, we novelly incorporate the AUC (area under the curve) score into the optimizing objective function. To address the issue that deep probabilistic graphical models requires large number of labeled training samples, we propose to use simulated RNA-seq datasets to train our model. Through extensive experimental studies on both simulation datasets of two species and biological datasets, we show that DeepBound consistently and significantly outperforms the two existing methods.

show abstract

“…In the training set, true β contacts were calculated following the DSSP definition with isolated β -bridge pairs ignored. The DSSP assignment was simplified into 3 categories: H, E and C. The secondary structure probabilities were predicted by DeepCNF [35]. The MSAs were built by HHblits [41] against the UniProt20 database [42], from which residue contact maps were then predicted by CCMpred.…”

Section: Datasetmentioning

confidence: 99%

“…Besides ridge features, general properties of the input contact map and position of the target residue pair within the map are abstracted as map property features and position features, respectively. The predicted secondary structure probabilities (from DeepCNF [35]) are incorporated as additional features. All features are fed into a 3-stage random forest framework to predict residue pairing in interacting β strands.…”

Section: Brief Introduction Of the Modelmentioning

confidence: 99%

Identification of residue pairing in interacting β-strands from a predicted residue contact map

Mao

Wang

Zhang

et al. 2017

Preprint

View full text Add to dashboard Cite

Despite the rapid progress of protein residue contact prediction, predicted residue contact maps frequently contain many errors. However, information of residue pairing in β strands could be extracted from a noisy contact map, due to the presence of characteristic contact patterns in β -β interactions. This information may benefit the tertiary structure prediction of mainly β proteins. In this work, we introduce a novel ridge-detection-based β -β contact predictor, RDb 2 C, to identify residue pairing in β strands from any predicted residue contact map. The algorithm adopts ridge detection, a well-developed technique in computer image processing, to capture consecutive residue contacts, and then utilizes a novel multi-stage random forest framework to integrate the ridge information and additional features for prediction. Starting from the predicted contact map of CCMpred, RDb 2 C remarkably outperforms all state-of-the-art methods on two conventional test sets of β proteins (BetaSheet916 and BetaSheet1452), and achieves F1-scores of ~62% and ~76%at the residue level and strand level, respectively. Taking the prediction of the more advanced RaptorX-Contact as input, RDb 2 C achieves impressively higher performance, with F1-scores reaching ~76% and ~86% at the residue level and strand level, respectively. According to our tests on 61 mainly β proteins, improvement in the β -β contact prediction can further ameliorate the structural prediction. Availability:All source data and codes are available at http://166.111.152.91/Downloads.html or at the GitHub address of https://github.com/wzmao/RDb2C. Author summaryDue to the topological complexity, mainly β proteins are challenging targets in protein structure prediction. Knowledge of the pairing between β strands, especially the residue pairing pattern, can greatly facilitate the tertiary structure prediction of mainly β proteins. In this work, we developed a novel algorithm to identify the residue pairing in β strands from a predicted residue contact map.This method adopts the ridge detection technique to capture the characteristic pattern of β -β interactions from the map and then utilizes a multi-stage random forest framework to predict β -β contacts at the residue level. According to our tests, our method could effectively improve the prediction of β -β contacts even from a highly noisy contact map. Moreover, the refined β -β contact information could effectively improve the structural modeling of mainly β proteins.

show abstract

DeepCNF-D: Predicting Protein Order/Disorder Regions by Weighted Deep Convolutional Neural Fields

Cited by 68 publications

References 43 publications

Evaluation of network architecture and data augmentation methods for deep learning in chemogenomics

Evaluation of network architecture and data augmentation methods for deep learning in chemogenomics

DeepBound: Accurate Identification of Transcript Boundaries via Deep Convolutional Neural Fields

Identification of residue pairing in interacting β-strands from a predicted residue contact map

Contact Info

Product

Resources

About