2017
DOI: 10.1101/103994
|View full text |Cite
Preprint
|
Sign up to set email alerts
|

Deep Recurrent Neural Network for Protein Function Prediction from Sequence

Abstract: As highthroughput biological sequencing becomes faster and cheaper, the need to extract useful information from sequencing becomes ever more paramount, often limited by lowthroughput experimental characterizations. For proteins, accurate prediction of their functions directly from their primary aminoacid sequences has been a long standing challenge. Here, machine learning using artificial recurrent neural networks (RNN) was applied towards classification of protein function directly from primary sequence witho… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2

Citation Types

0
44
0

Year Published

2018
2018
2024
2024

Publication Types

Select...
3
3
3

Relationship

0
9

Authors

Journals

citations
Cited by 67 publications
(44 citation statements)
references
References 20 publications
0
44
0
Order By: Relevance
“…Due to protein primary sequences' nature of position-specific residues dictating function, they are strong candidates for the application of LSTM networks as well. This potential is illustrated by a recent study which found that LSTM-based RNNs applied to protein sequences from UniProt were able to accurately predict several basic functions [8]. However, the computational methods employed in that study had several limitations.…”
Section: Introductionmentioning
confidence: 99%
“…Due to protein primary sequences' nature of position-specific residues dictating function, they are strong candidates for the application of LSTM networks as well. This potential is illustrated by a recent study which found that LSTM-based RNNs applied to protein sequences from UniProt were able to accurately predict several basic functions [8]. However, the computational methods employed in that study had several limitations.…”
Section: Introductionmentioning
confidence: 99%
“…In [24] a graph convolutional network reduces the required family size to 20 sequences, reporting 58% accuracy. It is encouraging that novel deep learning predictions for four functional classes were experimentally validated in [25].…”
mentioning
confidence: 96%
“…In [24] a graph convolutional network reduces the required family size to 20 sequences, reporting 58% accuracy. It is encouraging that novel deep learning predictions for four functional classes were experimentally validated in [25].Building confidence in deep learning approaches requires benchmarks that enable fair and rigorous comparison with existing state of the art methods and among deep model architectures. In this paper we pose protein function prediction as a sequence annotation task.…”
mentioning
confidence: 99%
“…Recently, this flexibility was demonstrated in protein structure prediction, replacing complex informatics pipelines with models that can predict structure directly from sequence 16 . Additionally, deep learning has shown success in sub-problems of protein informatics; for example: variant effect prediction 15 , function annotation 17,18 , semantic search 18 , and model-guided protein engineering 3,4 . While exciting advances, these methods are domain-specific or constrained by data scarcity due to the high cost of protein characterization.…”
mentioning
confidence: 99%