Table 2_OFGPMA: Optimal frequency graph representation learning for pseudogene and miRNA association prediction.docx
<p>Pseudogenes are genomic segments that resemble functional genes structurally yet remain biologically inactive. MicroRNAs (miRNAs), a subclass of non-coding RNAs, are critical regulators of various cellular mechanisms. These pseudogenes and miRNAs interact mutually, forming competitive endog...
Saved in:
| Hovedforfatter: | |
|---|---|
| Andre forfattere: | , |
| Udgivet: |
2025
|
| Fag: | |
| Tags: |
Tilføj Tag
Ingen Tags, Vær først til at tagge denne postø!
|
| Summary: | <p>Pseudogenes are genomic segments that resemble functional genes structurally yet remain biologically inactive. MicroRNAs (miRNAs), a subclass of non-coding RNAs, are critical regulators of various cellular mechanisms. These pseudogenes and miRNAs interact mutually, forming competitive endogenous RNA (ceRNA) networks alongside mRNA to influence physiological processes. Such regulatory networks have been implicated in numerous pathological conditions. Consequently, investigating pseudogene-miRNA associations holds promise for advancing disease diagnostics. Nevertheless, existing approaches to identify these relationships predominantly rely on labor-intensive experimental techniques, demanding substantial time and financial investments. Consequently, developing an effective computational framework that can identify new pseudogene-miRNA associations (PMAs) is crucial. To this end, we propose an optimal frequency graph representation learning framework named OFGPMA, for pseudogene-miRNA association prediction. OFGPMA enhances graph neural network expressiveness by learning both high-frequency energy and low-frequency energy components within the pseudogene-miRNA bipartite graph, utilizing Rayleigh and Chebyshev pooling techniques. This approach captures the graph’s global topology via Random Walk with Restart (RWR) and identifies potential local substructure features through enclosing subgraph analysis, thereby achieving a more comprehensive integration of the entire graph information. Comprehensive experiments show that OFGPMA outperforms state-of-the-art methods in terms of performance, while also exhibiting excellent generalization capabilities.</p> |
|---|