Kernel Cox regression models for linking gene expression profiles to censored survival data.

Area of Science:

• Genomics
• Bioinformatics
• Biostatistics

Background:

• Relating gene expression profiles to clinical phenotypes is crucial in functional genomics.
• Previous research focused on molecular classification of cancer, with less attention to censored survival data like patient survival or relapse times.

Purpose of the Study:

• To develop a novel kernel Cox regression model for analyzing gene expression data and censored survival phenotypes.
• To address the challenge of censored data in linking gene expression to patient outcomes.

Main Methods:

• A penalization method within reproducing kernel Hilbert spaces was employed.
• Negative partial likelihood was used as a loss function to handle censored data.
• The model identifies functional combinations of gene expression data correlated with survival times.

Main Results:

• The developed method successfully identified subgroups of patients with distinct risks of death or relapse.
• The identified gene expression patterns were highly correlated with patient survival times.
• The model accounted for variability in gene expression levels.

Conclusions:

• The kernel Cox regression model is effective for predicting relapse or death risk using gene expression profiles.
• The method demonstrates utility in analyzing censored survival data in diverse cancer types, including lymphoma and lung and breast carcinomas.