Codon usage signatures and codon pair usage in genes implicated in Stroke

Abstract

Ischaemic strokes account for the majority of strokes, which are a leading cause of death and permanent disability globally. In order to identify molecular patterns connected to disease-specific gene expression, this study investigates codon use bias in stroke-associated genes. With the aid of statistical testing and machine learning models, we examined nucleotide composition, codon usage indices, and Relative Synonymous Codon Usage (RSCU) using genomic data from BrainBase and HRT Atlas.The findings showed a strong bias in favour of GC-rich codons, especially at the third codon position, indicating adaptive selection for improved translational efficiency and mRNA stability under stress. Using only codon characteristics, machine learning classifiers—Random Forest in particular—were able to differentiate between genes linked to stroke and those involved in housekeeping. These results point to codon use bias as a possible molecular indicator of stroke, providing encouraging paths for the identification of biomarkers and better genomic categorisation.