About: Unsupervised protein embeddings outperform hand-crafted sequence and structure features at predicting molecular function

Facets (new session)
Description
Metadata
Settings
- owl:sameAs
- Inference Rule:

About: Unsupervised protein embeddings outperform hand-crafted sequence and structure features at predicting molecular function Goto Sponge NotDistinct Permalink

An Entity of Type : schema:ScholarlyArticle, within Data Space : wasabi.inria.fr associated with source document(s)

Attributes	Values
type	Academic Article research paper schema:ScholarlyArticle
isDefinedBy	Covid-on-the-Web dataset
has title	Unsupervised protein embeddings outperform hand-crafted sequence and structure features at predicting molecular function
Creator	Gomez, Angel Makrodimitris, Stavros Reinders, Marcel Sanchez, Victoria Van Ham, Roeland Villegas-Morcillo, Amelia
Source	BioRxiv
abstract	Motivation Protein function prediction is a difficult bioinformatics problem. Many recent methods use deep neural networks to learn complex sequence representations and predict function from these. Deep supervised models require a lot of labeled training data which are not available for this task. However, a very large amount of protein sequences without functional labels is available. Results We applied an existing deep sequence model that had been pre-trained in an unsupervised setting on the supervised task of protein function prediction. We found that this complex feature representation is effective for this task, outperforming hand-crafted features such as one-hot encoding of amino acids, k-mer counts, secondary structure and backbone angles. Also, it partly negates the need for deep prediction models, as a two-layer perceptron was enough to achieve state-of-the-art performance in the third Critical Assessment of Functional Annotation benchmark. We also show that combining this sequence representation with protein 3D structure information does not lead to performance improvement, hinting that three-dimensional structure is also potentially learned during the unsupervised pre-training. Availability Implementations of all used models can be found at https://github.com/stamakro/GCN-for-Structure-and-Function. Contact ameliavm@ugr.es Supplementary information Supplementary data are available online.
has issue date	2020-04-08(xsd:dateTime)
bibo:doi	10.1101/2020.04.07.028373
has license	biorxiv
sha1sum (hex)	e54f0a964e8e2f9428c58fca1149093481340a1f
schema:url	https://doi.org/10.1101/2020.04.07.028373
resource representing a document's title	Unsupervised protein embeddings outperform hand-crafted sequence and structure features at predicting molecular function
schema:publication	bioRxiv
resource representing a document's body	covid:e54f0a964e8e2f9428c58fca1149093481340a1f#body_text
is schema:about of	named entity 'Assessment' named entity 'training data' named entity 'unsupervised' named entity 'Results' named entity 'Protein function prediction' named entity 'XXXXX' named entity 'ENOUGH' named entity 'DIFFICULT' named entity 'REPRESENTATION' named entity 'OF-' named entity 'ENCODING' named entity 'PERFORMANCE' named entity 'COM' named entity 'THESE' named entity 'performance improvement' named entity 'sequence' named entity 'learn' named entity 'sequence' named entity 'sequences' named entity 'counts' named entity 'Critical' named entity 'labeled' named entity 'lead' named entity 'setting' named entity 'one-hot' named entity 'functional' named entity 'deep neural networks' named entity 'Critical Assessment of Functional Annotation' named entity 'pre-trained' named entity 'one-hot' named entity 'one-hot' named entity 'Euclidean distance' named entity 'contact map' named entity 'neural network' named entity 'amino acid' named entity 'logistic regression' named entity 'amino acid' named entity 'CNNs' named entity 'amino acids' named entity 'Gene Ontology' named entity 'amino acid' named entity 'Gene Ontology' named entity 'amino acid sequence' named entity 'SwissProt' named entity 'amino acid' named entity 'secondary structure' named entity 'protein' named entity 'one-hot' named entity 'data type' named entity 'k-mer' named entity 'alpha carbon' named entity 'secondary structure' named entity 'amino acids' named entity 'contact map' named entity 'kmer' named entity 'protein sequences' named entity 'protein' named entity '3D structure' named entity 'hyperparameters' named entity 'amino acid' named entity 'logistic regression' named entity 'SVD' named entity 'nearest neighbor' named entity 'undirected graph' named entity 'protein' named entity 'Gene Ontology' named entity 'contact maps' named entity 'binary image' named entity 'sigmoid function' named entity 'process variable' named entity 'random graph' named entity 'reduce the dimensionality' named entity 'probability' named entity '3D structure' named entity 'amino acids' named entity 'one-hot' named entity 'Ontology' named entity 'k-NN classifier' named entity 'protein'

Faceted Search & Find service v1.13.91 as of Mar 24 2020

Alternative Linked Data Documents: Sponger | ODE Content Formats:

RDF

ODATA

Microdata

About

OpenLink Virtuoso version 07.20.3229 as of Jul 10 2020, on Linux (x86_64-pc-linux-gnu), Single-Server Edition (94 GB total memory)
Data on this page belongs to its respective rights holders.
Virtuoso Faceted Browser Copyright © 2009-2024 OpenLink Software