Association mapping of complex diseases with ancestral recombination graphs: models and efficient algorithms

Area of Science:

• Genetics
• Bioinformatics
• Computational Biology

Background:

• Association mapping aims to identify genes influencing traits but is limited by not fully utilizing genealogical histories.
• Ancestral Recombination Graphs (ARGs) represent these histories but are complex to integrate into practical mapping methods.

Purpose of the Study:

• To develop novel association mapping methods that explicitly construct and sample Ancestral Recombination Graphs (ARGs).
• To improve the efficiency and accuracy of gene mapping for complex and Mendelian traits.

Main Methods:

• Developed a method to sample minimum ARGs (minARGs) uniformly at random for moderate datasets.
• Created a faster ARG sampling method for larger datasets, sampling from a defined ARG subspace.
• Extended the "phenotype likelihood" problem for improved mapping.

Main Results:

• Empirical results demonstrate significant speed improvements in association mapping.
• Achieved definite, though sometimes small, improvements in mapping accuracy.
• The developed methods are practical for both moderate and larger datasets.

Conclusions:

• The new ARG sampling methods offer substantial speed-ups for genome-wide scans.
• These advancements enhance the ability to locate causative mutations more efficiently.
• The methods provide a practical approach to leverage genealogical information for gene mapping.