New Gene Embedding Learned from Biomedical Literature and Its Application in Identifying Cancer Drivers
Liang, L.; Lu, X.; Lu, S.
Show abstract
To investigate molecular mechanism of diseases, we need to understand how genes are functionally associated. Computational researchers have tried to capture functional relationships among genes by constructing an embedding space of genes from multiple sources of high-throughput data. However, correlations in high-throughput data does not necessarily imply functional relations. In this study, we generated gene embedding from literature by constructing semantic representation for each gene. This approach enabled us to cover genes less mentioned in literature and revealed novel functional relationships among genes. Evaluation showed that the learned gene embedding was consistent with pathway knowledge and enhanced the search for cancer driver genes. We further applied our gene embedding to identify protein complexes and functional modules from gene networks. Performance in both scenarios was significantly improved with gene embedding.
Matching journals
The top 5 journals account for 50% of the predicted probability mass.
Similar papers in this journal
- Novel cancer subtyping method based on patient-specific gene regulatory network 95%
- Discovering Key Transcriptomic Regulators in Pancreatic Ductal Adenocarcinoma using Dirichlet Process Gaussian Mixture Model 94%
- A Convolution Based Computational Approach Towards DNA N6-methyladenine Site Identification and Motif Extraction in Rice Genome 94%
Similar papers in this journal
- C3: Connect separate Connected Components to form a succinct disease module 96%
- Investigate the relevance of major signaling pathways in cancer survival using a biologically meaningful deep learning model 94%
- Multi-Head Attention-based U-Nets for Predicting Protein Domain Boundaries Using 1D Sequence Features and 2D Distance Maps 94%
Similar papers in this journal
- MeSHHeading2vec: A new method for representing MeSH headings as feature vectors based on graph embedding algorithm 95%
- Integrative COVID-19 Biological Network Inference with Probabilistic Core Decomposition 95%
- LSTM-PHV: Prediction of human-virus protein-protein interactions by LSTM with word2vec 94%
Similar papers in this journal
- Integration of multi-source gene interaction networks and omics data with graph attention networks to identify novel disease genes 96%
- 3Cnet: Pathogenicity prediction of human variants using knowledge transfer with deep recurrent neural networks 95%
- Multi-omics Data Integration by Generative Adversarial Network 94%
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.