Operationalizing language-based population stratification for widening access to precision genomics in Africa.

Area of Science:

• Population Genetics
• Computational Linguistics
• Genomic Diversity

Background:

• African genetic diversity is significantly under-represented in genomic studies (2.4%).
• This disparity limits understanding of human biology and equitable healthcare translation.
• Genomic data gaps hinder research in diverse populations.

Purpose of the Study:

• To evaluate lexical similarity as a proxy for genetic relationships in Kenya.
• To compare lexical distance with geographic proximity in predicting population structure.
• To provide a data-driven method for prioritizing populations in genomic research.

Main Methods:

• Utilized normalized Levenshtein distance (LDN) for lexical similarity analysis.
• Analyzed patterns across Kenya's Bantu, Nilotic, and Cushitic language groups.
• Compared lexical distance matrices with genetic differentiation and geographic data.

Main Results:

• Lexical similarity revealed distinct clusters mirroring ethnolinguistic diversity.
• Lexical distance strongly correlated with genetic differentiation (r=0.91).
• Language proved a superior proxy for population genetic structure compared to geography (r=0.29).

Conclusions:

• Lexical similarity analysis is a robust method for predicting genetic relationships in African populations.
• This approach aids in strategically prioritizing populations for genomic sequencing.
• Offers a systematic solution to enhance African genetic diversity in global databases.