Machine Learning Gene Signature to Metastatic ccRCC Based on ceRNA Network

Area of Science:

• Oncology
• Genomics
• Molecular Biology

Background:

• Clear-cell renal-cell carcinoma (ccRCC) is a significant pathology characterized by high metastatic rates.
• While coding genes' roles in metastasis are known, non-coding genes, like competitive endogenous RNA (ceRNA), are increasingly studied.
• Understanding the ceRNA network is crucial for identifying new biomarkers and therapeutic targets in ccRCC.

Purpose of the Study:

• To construct a ceRNA network for ccRCC associated with metastatic development.
• To develop and validate a gene signature for predicting ccRCC metastasis.
• To analyze the biological functions and prognostic significance of the identified gene signature.

Main Methods:

• Utilized The Cancer Genome Atlas (TCGA) and International Cancer Genome Consortium (ICGC-RECA) datasets.
• Constructed a ceRNA network using differentially expressed genes.
• Employed eight feature selection techniques to assemble and select a final 11-gene signature.
• Performed genomic, risk, and functional annotation analyses.

Main Results:

• An 11-gene signature (SNHG15, AF117829.1, hsa-miR-130a-3p, hsa-mir-381-3p, BTBD11, INSR, HECW2, RFLNB, PTTG1, HMMR, RASD1) was identified.
• The signature demonstrated good generalization with an Area Under the Curve (AUC) of 81.5% on an external dataset.
• Key genes (hsa-miR-130a-3p, AF117829.1, hsa-miR-381-3p, PTTG1) correlated significantly with patient survival and metastasis.
• Functional annotation revealed involvement in RNA polymerase II transcription regulation and cell cycle control.

Conclusions:

• The identified 11-gene signature, comprising coding and non-coding genes, shows promise as a biomarker for ccRCC.
• The ceRNA network analysis suggests lncRNAs acting as sponges for miRNAs, contributing to ccRCC progression.
• This signature aids in understanding ccRCC biology and could improve patient prognosis and management.