sNMF

sNMF estimates individual ancestry coefficients from genotype matrices using sparse non-negative matrix factorization for population genetic and association studies.

Key Features:

• Sparse Non-negative Matrix Factorization: Employs a sparse NMF algorithm to decompose genotype matrices into interpretable ancestry components representing individual ancestry proportions.
• Efficiency and Speed: Achieves runtimes approximately 10–30 times faster than traditional likelihood-based methods such as ADMIXTURE while maintaining high accuracy.
• Implementation: Implemented as a computational program for rapid estimation of ancestry coefficients suitable for large-scale genomic datasets.

Scientific Applications:

• Population Genetics: Provides fast and accurate estimates of individual ancestry coefficients to investigate genetic structure and admixture in human populations.
• Association Studies: Accounts for population stratification by estimating ancestry proportions to improve identification of genotype–phenotype associations.
• Cross-Species Analysis: Has been applied to plant genomic datasets, demonstrating applicability beyond human data.

Methodology:

sNMF leverages sparse non-negative matrix factorization to decompose genotype matrices into ancestry components and infer individual ancestry coefficients as a computationally efficient alternative to likelihood-based approaches such as ADMIXTURE.