What question did this study set out to answer?

To develop and evaluate a deep learning model that accurately predicts drug-target interactions directly from raw protein sequences by capturing local residue patterns.

June 14, 2019PLoS Computational BiologyOpen Access

DeepConv-DTI: Prediction of drug-target interactions via deep learning with convolution on protein sequences

View Full Paper

Ask AI

Bookmark

Authors

ILIngoo LeeUniversity of California, San Diego JKJongsoo KeumGwangju Institute of Science and Technology HNHojung NamGwangju Institute of Science and Technology

Discussion

Member takes

Implication

Computational modeling study demonstrates improved drug-target interaction prediction across diverse proteins, highlighting deep learning's ability to detect functional binding sites.

Key Points

To develop and evaluate a deep learning model that accurately predicts drug-target interactions directly from raw protein sequences by capturing local residue patterns.
Designed a convolutional neural network (CNN) framework, DeepConv-DTI, that applies convolution filters across amino acid subsequences of various lengths.
Trained the architecture on large-scale drug-target interaction records and benchmarked performance against an independent, previously unseen dataset.
DeepConv-DTI achieved superior predictive accuracy compared to traditional descriptor-based models and existing deep learning architectures on the independent validation set.
Analysis of pooled convolution features demonstrated that the model effectively localized functional binding sites on target proteins without requiring three-dimensional structural data.

Cite This Study

Lee et al. (2019) studied this question.

synapsesocial.com/papers/69d8970cde3177251abed9d2 https://doi.org/10.1371/journal.pcbi.1007129

View Full Paper

Ask AI

Bookmark

Implication

Computational modeling study demonstrates improved drug-target interaction prediction across diverse proteins, highlighting deep learning's ability to detect functional binding sites.

DeepConv-DTI: Prediction of drug-target interactions via deep learning with convolution on protein sequences

Discussion

Member takes

Implication

Key Points

Cite This Study

DeepConv-DTI: Prediction of drug-target interactions via deep learning with convolution on protein sequences

Discussion

Member takes

Implication

Key Points

Cite This Study

Also Consider

Also Consider