SpartaABC

SpartaABC infers insertion and deletion (indel) parameters from multiple sequence alignments using an approximate Bayesian computation (ABC) rejection algorithm to enable realistic sequence simulations.

Key Features:

• Indel parameter inference: Extracts summary statistics from multiple sequence alignments and applies an ABC rejection algorithm that samples sequences under randomly chosen indel parameters and retains parameter sets whose simulated statistics match the input.
• Output of inferred parameters: Produces point estimates and approximate posterior distributions for inferred indel parameters.
• Sequence simulation integration: Uses Dawg 2.0 and INDELible to generate simulated sequences under inferred indel parameters.
• Distance-based model selection: Computes distances between summary statistics from input sequence data and each simulation to identify parameter sets that closely mimic the empirical data.

Scientific Applications:

• Evolutionary analysis of indel dynamics: Enables quantitative inference of indel rates and length distributions for studies of sequence evolution.
• Comparative genomics: Facilitates comparison of indel dynamics across genes or species, including analyses of orthologous protein-coding genes.
• Simulation-based hypothesis testing: Allows generation of simulated datasets under inferred indel parameters to compare against empirical observations for model validation.

Methodology:

SpartaABC extracts summary statistics from input sequence data, performs numerous simulations with randomly sampled indel parameters using an ABC rejection algorithm, computes distances between observed and simulated summary statistics to retain matching parameter sets, and reports inferred indel parameters while enabling further sequence simulations with Dawg 2.0 or INDELible.