protein2vec: Predicting Protein-Protein Interactions Based on LSTM

Zhang, Jiongmin; Zhu, Man; Qian, Ying

doi:10.1109/tcbb.2020.3003941

Cited by 13 publications

(13 citation statements)

References 36 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Multiple studies approached the prediction of PPIs using various sources of information, such as the primary structure of the protein ( Chen et al , 2019 ; Hashemifar et al , 2018 ; Li et al , 2018 ), the 3D protein structure ( Bepler and Berger, 2021 ), gene expression profiles ( Chin et al , 2010 ) and Gene Ontology (GO) annotation ( Bandyopadhyay and Mallick, 2017 ; Jain and Bader, 2010 ; Kulmanov et al , 2019 ; Smaili et al , 2018 , 2019 ; Zhang et al , 2018 , 2020 ).…”

Section: Introductionmentioning

confidence: 99%

“…Disregarding the structure of the ontology would not allow for a correct evaluation of proteins that have common terms but are too general ( Guo et al , 2006 ). Several studies apply techniques from the field of Natural Language Processing to extract dense feature vectors for GO terms ( Smaili et al , 2018 , 2019 ; Zhang et al , 2020 ; Zhao et al , 2020 ; Zhong et al , 2019 ). We find that previous work comparing feature vectors using cosine similarity or using a fully connected neural network fail to capture deep semantic similarity between the GO terms.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

TransformerGO: predicting protein–protein interactions by modelling the attention between sets of gene ontology terms

2022

View full text Add to dashboard Cite

Motivation Protein-protein interactions (PPIs) play a key role in diverse biological processes but only a small subset of the interactions have been experimentally identified. Additionally, high-throughput experimental techniques that detect PPIs are known to suffer various limitations such as exaggerated false positives and negatives rates. The semantic similarity derived from the Gene Ontology (GO) annotation is regarded as one of the most powerful indicators for protein interactions. However, while computational approaches for prediction of PPIs have gained popularity in recent years, most methods fail to capture the specificity of GO terms. Results We propose TransformerGO, a model that is capable of capturing the semantic similarity between gene ontology sets dynamically using an attention mechanism. We generate dense graph embeddings for GO terms using an algorithmic framework for learning continuous representations of nodes in networks called node2vec. TransformerGO learns deep semantic relations between annotated terms and can distinguish between negative and positive interactions with high accuracy. TransformerGO outperforms classic semantic similarity measures on gold standard PPI datasets and state-of-the-art machine learning-based approaches on large datasets from S. cerevisiae and H. sapiens. We show how the neural attention mechanism embedded in the transformer architecture detects relevant functional terms when predicting interactions. Availability and implementation https://github.com/Ieremie/TransformerGO Supplementary information Supplementary data are available at Bioinformatics online.

show abstract

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

TransformerGO: predicting protein–protein interactions by modelling the attention between sets of gene ontology terms

2022

View full text Add to dashboard Cite

show abstract

“…While research exploring the application of these methods to biological networks is still in its early stages, there is considerable interest. These applications can be loosely divided into three categories; (1) drug-related applications , such as drug-target interactions (DTIs) (4–9), drug-disease interactions (10–12), drug side-effects (13, 14), drug-drug interactions (15–17), polypharmacy antagonistic effects (18, 19) and synergistic reactions in drug combination therapy (20); (2) protein-related applications , such as protein-protein interactions (PPIs) (21–24) and protein/gene disease interactions (25–31); and (3) transcriptomics-related applications , such as lncRNAs-diseases associations (32–35) and miRNAdisease associations (36–43) and many other applications (44–50).…”

Section: Introductionmentioning

confidence: 99%

“…(2) protein-related applications, such as protein-protein interactions (PPIs) (21)(22)(23)(24) and protein/gene disease interactions (25)(26)(27)(28)(29)(30)(31); and (3) transcriptomics-related applications, such as lncRNAs-diseases associations (32)(33)(34)(35) and miRNAdisease associations (36)(37)(38)(39)(40)(41)(42)(43) and many other applications (44)(45)(46)(47)(48)(49)(50). Since network embedding methods were not originally developed for biological networks, their performance in obtaining different biological network features is yet to be established.…”

Section: Introductionmentioning

confidence: 99%

BioNE: Integration of network embeddings for supervised learning

Parvizi

Azuaje

Τheodoratou

et al. 2022

Preprint

View full text Add to dashboard Cite

SummaryA network embedding approach reduces the analysis complexity of large biological networks by converting them to lowdimensional vector representations (features/embeddings). These lower-dimensional vectors can then be used in machine learning prediction tasks with a wide range of applications in computational biology and bioinformatics. Several network embedding approaches have been proposed with different methods of generating vector representations. These network embedding approaches can be quite diverse in terms of data representation and implementation. Moreover, most were not originally developed for biological networks. Therefore comparing and assessing the performance of these diverse models in practice, in biological contexts, can be challenging. To facilitate such comparisons, we have developed the BioNE framework for integration of different embedding methods in prediction tasks. Using this framework one can easily assess, for instance, whether combined vector representations from multiple embedding methods offer complementary information with regards to the network features and thus better performance on prediction tasks. In this paper, we present the BioNE software suite for embedding integration, which applies network embedding methods following standardised network preparation steps, and integrates the vector representations achieved by these methods using three different techniques. BioNE enables selection of prediction models, oversampling methods, feature selection methods, cross-validation type and cross-validation parameters.Availability and implementationBioNE pipeline and detailed explanation of implementation is freely available on GitHub, at https://github.com/pooryaparvizi/BioNE

show abstract

“…protein protein interactions [79], and relation prediction for ontology population [41]. Here, we used both protein ontology (PRO) and Gene Ontology (GO) to detect inconsistencies and conflicting functions among the list of annotations assigned to a single protein sequence.…”

Section: Graph Embeddingmentioning

confidence: 99%

Towards data cleaning in large public biological databases

Bagheri¹

View full text Add to dashboard Cite

show abstract

protein2vec: Predicting Protein-Protein Interactions Based on LSTM

Cited by 13 publications

References 36 publications

TransformerGO: predicting protein–protein interactions by modelling the attention between sets of gene ontology terms

TransformerGO: predicting protein–protein interactions by modelling the attention between sets of gene ontology terms

BioNE: Integration of network embeddings for supervised learning

Towards data cleaning in large public biological databases

Contact Info

Product

Resources

About