Bayesian genome assembly and assessment by markov chain monte carlo sampling

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Traditional genome assemblers produce single sequence estimates, overlooking data-supported alternative hypotheses.
• Evaluating the reliability and uncertainty of genome assemblies remains a challenge.

Purpose of the Study:

• To develop a novel Markov chain Monte Carlo (MCMC) approach for sequence assembly that generates probabilistic distributions of assembly hypotheses.
• To provide a robust statistical framework for assessing alternative genome sequences and quantifying assembly uncertainty.

Main Methods:

• Implementation of a Bayesian inference framework using MCMC sampling.
• Development of a prototype assembler named Genome Assembly by Bayesian Inference (GABI).
• Application and validation using bacteriophage φX174 genome data and Illumina sequencing reads.

Main Results:

• Demonstrated feasibility of the MCMC approach with good mixing and convergence on test data.
• Generated posterior distributions of assembly hypotheses, summarized as a majority-rule consensus assembly.
• Enabled annotation of external assemblies by assigning posterior probabilities to common features.

Conclusions:

• The Bayesian MCMC approach offers a statistically rigorous method for genome assembly.
• GABI provides a powerful tool for evaluating assembly uncertainty and alternative hypotheses.
• This method advances the field of computational genomics by incorporating probabilistic reasoning.