Predicting Environmental and Ecological Drivers of Human Population Structure

Area of Science:

• Population Genetics
• Human Evolution
• Machine Learning Applications

Background:

• Understanding factors shaping human genetic diversity is crucial.
• Existing methods struggle to analyze multiple environmental and cultural variables simultaneously.
• East Africa's rich diversity provides a unique setting to study population structure.

Purpose of the Study:

• To develop and apply a machine learning method for identifying key drivers of human migration rates.
• To disentangle the influence of landscape, climate, and tsetse fly presence on genetic patterns.
• To investigate adaptation signatures in Ethiopian populations.

Main Methods:

• Utilized a novel machine learning approach to analyze migration rates inferred by the MAPS program.
• Applied the method to high-density single nucleotide polymorphism data from 30 East African populations.
• Incorporated over 20 spatial variables including landscape, climate, and tsetse fly distribution.

Main Results:

• The model explained approximately 40% of the variance in human migration rates over the past 56 generations.
• Precipitation, minimum temperature of the coldest month, and elevation were identified as the most impactful environmental variables.
• The fusca tsetse fly group, a vector for trypanosomiasis, also showed significant impact.
• Signatures of positive selection related to metabolism and disease were found in Ethiopian high-elevation populations, but not well-known high-elevation genes.

Conclusions:

• Environmental factors play a significant role in shaping human migration and genetic adaptation in East Africa.
• The study highlights the power of machine learning in dissecting complex population genetic patterns.
• Unexplained genetic variation suggests the influence of unmeasured cultural or other factors.