LiteQTL

LiteQTL performs whole-genome quantitative trait locus (QTL) scans to accelerate detection of genotype-phenotype associations in large-scale omics datasets using CPU and GPU parallelization.

Key Features:

• High-Speed Performance: Achieves speeds reported to be more than 700× faster than traditional R/qtl linear model genome scans when using 16 threads.
• GPU Acceleration: Harnesses GPUs and CPU multi-threading to accelerate computations, with performance dependent on problem size and shape (number of cases, genotypes, and traits).
• Scalability: Capable of handling up to one million traits in a single scan for large-scale genetic studies.
• Parallelizable Operations: Optimizes easily parallelizable computations such as matrix multiplication, vectorized operations, and element-wise computations.

Scientific Applications:

• Systems genetics and QTL mapping: Enables rapid whole-genome scans to identify genotype-phenotype correlations across large omics datasets.
• Reference and segregating population analysis: Supports analysis of reference populations such as the BXD family of mouse strains, which are fully sequenced and deeply phenotyped, for studying complex traits.

Methodology:

Implements parallelized algorithms using multi-threading and GPU acceleration to optimize matrix multiplication, vectorized operations, and element-wise computations; the implementation is provided in the Julia programming language.