ntRoot: computational inference of human ancestry at scale from genomic data

Area of Science:

• Genomics
• Population Genetics
• Bioinformatics

Background:

• Accurate ancestry information is crucial for large cohort studies.
• Existing ancestry prediction methods face computational and input complexity challenges with genome sequencing data.
• There is a need for efficient, scalable, and accurate methods for inferring ancestry directly from sequencing data.

Purpose of the Study:

• To present ntRoot, a computationally lightweight method for inferring human super-population-level ancestry.
• To enable rapid and resource-efficient ancestry inference from diverse genomic data types.
• To advance ancestry characterization in large-scale genomic research.

Main Methods:

• Developed ntRoot, a reference-guided, alignment-free single nucleotide variant detection framework.
• Utilized a Bloom filter for efficient querying of genomic inputs against a variant reference panel.
• Tested on over 600 human genome samples, including complete genomes, draft assemblies, and sequencing data.

Main Results:

• ntRoot accurately predicts geographic labels and shows high concordance with traditional methods like ADMIXTURE (R² = 0.9567) for ancestry fractions.
• Analysis completion times are efficient: 30 minutes for assemblies and 75 minutes for 30-fold sequencing data.
• The method provides both global and local ancestry inference with high-resolution predictions.

Conclusions:

• ntRoot addresses the critical need for scalable, accurate, and resource-efficient ancestry inference in cohort studies.
• The method offers a significant advancement for ancestry characterization in genomic research.
• ntRoot is freely available, promoting its adoption and utility.