CONE: Community Oriented Network Estimation Is a Versatile Framework for Inferring Population Structure in...

Area of Science:

• Population genetics
• Ecology
• Bioinformatics

Background:

• Estimating genetic population structure from molecular markers is crucial in population genetics and ecology.
• Conventional methods often have limitations regarding sample composition and pre-defined parameters.

Purpose of the Study:

• To develop a flexible and accurate method for inferring population genetic structure and gene flow.
• To introduce CONE (community oriented network estimation) as an advancement over existing techniques.

Main Methods:

• Application of a generalized linear model with LASSO regularization for relationship inference.
• Utilizing a neighborhood selection algorithm and community detection for network-based structure analysis.
• Construction of an individual-level population graph to visualize genetic relationships.

Main Results:

• CONE provided more accurate estimates of the true number of subpopulations compared to model-based methods on simulated data.
• Ancestry coefficient estimates from CONE were comparable to those from traditional methods.
• Analysis of empirical datasets (teosinte, bacterial outbreak, human genome diversity) showed CONE's results align with previous findings.

Conclusions:

• CONE offers a robust and flexible approach to population structure inference, overcoming limitations of conventional methods.
• The network-based approach of CONE combines strengths from network theory, PCA, and model-based techniques.
• CONE demonstrates broad applicability across diverse biological systems and data types.