Amino Acid Encoding for Deep Learning Applications

Background: The number of applications of deep learning algorithms in bioinformatics is increasing as they usually achieve superior performance over classical approaches, especially, when bigger training datasets are available. In deep learning applications, discrete data, e.g. words or n-grams in l...

Full description

Saved in:
Bibliographic Details
Published in:BMC bioinformatics Vol. 21; no. 1; pp. 235 - 14
Main Authors: ElAbd, Hesham, Bromberg, Yana, Hoarfrost, Adrienne, Lenz, Tobias, Wendorff, Mareike
Format: Journal Article
Language:English
Published: 2230 Support BMC 09.06.2020
BioMed Central
BioMed Central Ltd
Springer Nature B.V
Subjects:
Algorithms
Amino acid encoding
Amino acids
Amino Acids - metabolism
Amino acids embedding
Artificial neural networks
Bioinformatics
Biomedical and Life Sciences
Cable television broadcasting industry
Computational Biology - methods
Computational Biology/Bioinformatics
Computer Appl. in Life Sciences
Computer applications
Convoluted-neural network (CNN)
Cybernetics, Artificial Intelligence and Robotics
Data mining
Deep learning
Deep Learning - standards
Embedding
HLA-II peptide interaction
Humans
Iterative methods
Learning algorithms
Life Sciences
Machine learning
Machine Learning and Artificial Intelligence in Bioinformatics
Mathematical analysis
Matrix algebra
Matrix methods
Methodology
Methodology Article
Microarrays
Neural networks
Nucleotides
Observational learning
Peptides
Protein-protein interaction (PPI)
Protein-protein interactions
Proteins
Recurrent neural networks
Technology application
Training
United States
Recurrent Neural Network (rnn)
Human-Leukocyte Antigen (hla)
Amino Acids Embedding
Amino Acid Encoding
Hla-Ii Peptide Interaction
Deep-Learning
Convoluted-Neural Network (cnn)
Machine-Learning (ml)
Protein-Protein Interaction (ppi)
Protein-protein interaction (PPI)
HLA-II peptide interaction
Deep-learning
Convoluted-neural network (CNN)
Recurrent neural network (RNN)
Human-leukocyte antigen (HLA)
Machine-learning (ML)
Amino acid encoding
Amino acids embedding
ISSN:1471-2105, 1471-2105
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Background: The number of applications of deep learning algorithms in bioinformatics is increasing as they usually achieve superior performance over classical approaches, especially, when bigger training datasets are available. In deep learning applications, discrete data, e.g. words or n-grams in language, or amino acids or nucleotides in bioinformatics, are generally represented as a continuous vector through an embedding matrix. Recently, learning this embedding matrix directly from the data as part of the continuous iteration of the model to optimize the target prediction – a process called ‘end-to-end learning’ – has led to state-of-the-art results in many fields. Although usage of embeddings is well described in the bioinformatics literature, the potential of end-to-end learning for single amino acids, as compared to more classical manually-curated encoding strategies, has not been systematically addressed. To this end, we compared classical encoding matrices, namely one-hot, VHSE8 and BLOSUM62, to end-to-end learning of amino acid embeddings for two different prediction tasks using three widely used architectures, namely recurrent neural networks (RNN), convolutional neural networks (CNN), and the hybrid CNN-RNN. Results: By using different deep learning architectures, we show that end-to-end learning is on par with classical encodings for embeddings of the same dimension even when limited training data is available, and might allow for a reduction in the embedding dimension without performance loss, which is critical when deploying the models to devices with limited computational capacities. We found that the embedding dimension is a major factor in controlling the model performance. Surprisingly, we observed that deep learning models are capable of learning from random vectors of appropriate dimension. Conclusion: Our study shows that end-to-end learning is a flexible and powerful method for amino acid encoding. Further, due to the flexibility of deep learning systems, amino acid encoding schemes should be benchmarked against random vectors of the same dimension to disentangle the information content provided by the encoding scheme from the distinguishability effect provided by the scheme.
Bibliography:2230 Support
2230
ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
content type line 23
ISSN:1471-2105
1471-2105
DOI:10.1186/s12859-020-03546-x