Welcome to Cheeta v1.0!
Complete genome-wide SNP×SNP interaction scanning of biobank data within 24 hours.

Introduction

Genome-wide SNP×SNP interaction analysis provides a powerful opportunity to uncover epistatic mechanisms underlying complex diseases, yet its substantial computational demands – involving trillions of pairwise tests across hundreds of thousands of individuals and millions of loci – have long remained a prohibitive barrier for exhaustive scanning.

Therefore, we define cheeta (version 1.0), a GPU‑accelerated exhaustive epistasis scanning software for large‑scale biobanks (a single‑consumer‑GPU parallel computing tool for Windows). Cheeta is a user‑friendly command‑line tool that performs ultrafast, exhaustive genome‑wide SNP×SNP interaction analysis at hour‑level turnaround. The tool supports three complementary interaction models: a comprehensive nine‑genotype combination model for full pairwise genotype interaction detection, a dominant‑recessive framework consistent with classical Mendelian inheritance, and a multifactor dimensionality reduction approach that clusters genotypes into high‑ and low‑risk categories. Flexible output options allow users to either export all pairwise results or apply specific filtering criteria (e.g., odds ratio and corrected p‑value thresholds). Cheeta also provides optional global summary statistics of detected interactions. As a specialized and highly efficient solution for biobank‑scale exhaustive epistasis analysis, it eliminates the need for high‑performance computing clusters or specialized hardware, empowering researchers with standard GPU hardware to efficiently uncover previously inaccessible genetic architectures of complex diseases.