miqoGraph

miqoGraph applies mixed-integer quadratic programming to infer and fit admixture graphs that represent population splits, drift, and admixture for analysis of genetic relationships among populations.

Key Features:

• Mixed-integer quadratic programming (MIQP): Uses MIQP to optimize parameters of admixture graphs.
• Simultaneous parameter fitting: Simultaneously fits graph topology, drift lengths, and admixture proportions within a unified optimization framework.
• Automated topology inference: Infers admixture graph topology using integer optimization techniques.
• Explicit evolutionary modeling: Represents splits (population divergence), genetic drift, and admixture (gene flow) explicitly in the graph.
• Julia implementation: Implemented within the Julia package ecosystem.
• Scalability and computational feasibility: Targets accurate and computationally feasible solutions for large datasets and intricate evolutionary scenarios.

Scientific Applications:

• Population history reconstruction: Reconstructs complex population histories using admixture graphs that include splits, drift, and admixture.
• Parameter estimation: Estimates drift lengths and admixture proportions to quantify genetic change and gene flow.
• Topology discovery: Infers admixture graph topologies that explain observed genetic relationships among populations.
• Analysis of complex datasets: Analyzes large and intricate evolutionary scenarios involving multiple admixture events and drift.

Methodology:

miqoGraph applies mixed-integer quadratic programming (MIQP) and integer optimization to simultaneously fit admixture graph topology, drift lengths, and admixture proportions.