Decoupling between sickle cell disease and African ancestry in Brazil: Deep Learning as a tool for understanding admixture dynamics and decoupling

• Grant number: 24/23088-5
• Support Opportunities: Scholarships abroad - Research Internship - Scientific Initiation
• Start date: April 01, 2025
• End date: July 31, 2025
• Field of knowledge: Biological Sciences - Genetics
• Principal Investigator: Diogo Meyer
• Grantee: Pedro Nicésio do Amaral
• Supervisor: Matteo Fumagalli
• Host Institution: Instituto de Biociências (IB). Universidade de São Paulo (USP). São Paulo , SP, Brazil
• Institution abroad: Queen Mary University of London, England
• Associated to the scholarship: 24/05419-4 - Decoupling between sickle-cell disease and african ancestry in Brazil, BP.IC

Abstract

This project aims to explore the decoupling process between sickle cell disease (SCD) and African ancestry within the Brazilian population, with a focus on understanding how demographic factors such as assortative mating and sex bias shape genetic diversity in admixed populations. The research builds on findings from a FAPESP-funded undergraduate and Master's project, which revealed a significant degree of decoupling: many individuals with SCD in Brazil exhibit higher European than African ancestry. The proposed internship aims to advance these findings by integrating Deep Learning methods into demographic analyses. Specifically, the project will leverage neural networks to infer demographic parameters, enhancing the analytical power beyond that of traditional methods. By training deep neural networks on simulated data - generated through mechanistic models of assortative mating and sex bias - we will create predictive frameworks that reflect real-world admixture dynamics. Collaboration with Matteo Fumagalli's team is central to this initiative, providing expertise in applying Deep Learning to population genetics. The internship will also analyze IBGE records, offering longitudinal demographic data on marriage patterns across Brazilian states. This is a source of data which is orthogonal to that based in genetics, and will provide an independent perspective on the degree of assortative mating in Brazil. Integrating these records with genomic data from the REDS-III cohort will allow a nuanced reconstruction of mating processes over time, addressing a gap identified in Fumagalli's prior work: the need to model changes in assortative mating through historical periods. By the project's conclusion, we expect to develop a robust mechanistic model of mating in Brazil, enabling precise inferences about the strength and impact of assortative mating and sex bias on genetic structure. This work not only deepens our understanding on the decoupling of SCD and African ancestry, but also lays the groundwork for future simulations that could predict demographic shifts in subsequent generations.