CLUMPAK

CLUMPAK postprocesses results from model-based clustering programs (e.g., STRUCTURE) to identify consensus clustering solutions and align inferred clusters across K for multilocus genotype analyses.

Key Features:

• Automated postprocessing: Handles multiple independent runs for each fixed value of K produced by model-based clustering programs such as STRUCTURE.
• Identification of similar runs: Identifies sets of highly similar replicate runs and distinct modes using a Markov clustering algorithm applied to similarity matrices computed by Clumpp.
• Consensus solution generation: Generates consensus solutions for each identified mode in the solution space.
• Alignment across K values: Aligns inferred clusters across varying values of K to facilitate comparison of clustering results obtained at different K.
• Selection of optimal K: Implements methods for selecting an optimal number of clusters (K).
• Comparison across programs, models, and data subsets: Compares solutions derived from different programs, models, or data subsets.

Scientific Applications:

• Population genetics: Interpretation and comparison of population genetic structure from multilocus genotype data.
• Molecular ecology: Investigation of population-level genetic structure in ecological and conservation studies.
• Multilocus genotype analysis: Summarization and comparison of multilocus genotype clustering results across runs and K values.

Methodology:

Similarity matrices are computed by Clumpp, a Markov clustering algorithm is applied to identify distinct modes among replicate runs, consensus solutions are generated for each mode, and inferred clusters are aligned across K values.