BiTUGA: scalable prevalence-based unitig association testing for binary traits

Area of Science:

• Genomics
• Bioinformatics
• Population Genetics

Background:

• Sequence-based association studies face challenges in large, repetitive genomes.
• Reference-free k-mer methods are powerful but computationally intensive due to large matrices.
• Existing tools struggle with the scale of modern genomic datasets.

Purpose of the Study:

• To introduce BiTUGA, a novel bioinformatics pipeline for sequence-based association testing in large genomes.
• To overcome computational bottlenecks associated with k-mer analysis in repetitive genomes.
• To enable efficient analysis of discrete binary traits, such as sex determination.

Main Methods:

• BiTUGA utilizes a reference-free k-mer approach, focusing on unitig presence across sample groups.
• It shifts statistical analysis from raw k-mer abundance to group-level prevalence.
• The pipeline was validated using the binary trait of sex in plant species.

Main Results:

• BiTUGA successfully identified sex-associated unitigs in Populus tremula and Ginkgo biloba.
• The pipeline detected known Sex-Determining Regions (SDRs) in these species.
• BiTUGA processed large datasets (up to 750 Gbp) efficiently within 14-25 hours and 70 GB RAM.

Conclusions:

• BiTUGA offers an efficient and scalable solution for association studies in large genomes.
• The pipeline effectively identifies genetic regions associated with binary traits, like sex determination.
• BiTUGA provides a valuable tool for exploring complex genomic regions in various species.