BWA-PSSM

BWA-PSSM implements probabilistic short-read alignment using position-specific scoring matrices (PSSM) to improve mapping accuracy of short DNA sequencing reads to reference genomes.

Key Features:

• Position-Specific Scoring Matrices (PSSM): Uses PSSMs to score read-to-reference matches with position-dependent probabilities.
• Probabilistic mapping framework: Computes alignments within a probabilistic model rather than relying solely on simple mismatch counts.
• Integration of quality scores: Accounts for sequencing read quality scores in alignment scoring.
• Prior mismatch probabilities: Incorporates prior mismatch probabilities into alignment and scoring calculations.
• User-defined evolutionary and bias models: Accepts user-defined models of evolutionary processes and dataset-specific biases for adaptive alignment.
• Mapping quality computation: Produces mapping qualities derived from the underlying probabilistic model.
• Compatibility with BWA index: Implemented as an extension of BWA and retains compatibility with the BWA genome index structure.
• Improved sensitivity and accuracy: Demonstrated higher mapping quality than Bowtie and BWA on simulated and real datasets.
• Robustness to contaminants and close genomes: Mitigates random matches from short contaminant reads and improves reliability of mapping between closely related genomes.

Scientific Applications:

• Short-read alignment to reference genomes: Improves mapping accuracy for generic short DNA sequencing datasets.
• Ancient DNA analysis: Enhances mapping of degraded and low-quality ancient DNA reads.
• PAR-CLIP read mapping: Applies to mapping PAR-CLIP sequencing reads.
• AT-rich organism sequencing: Improves alignment for AT-rich genomes such as Plasmodium falciparum.
• Distinguishing closely related species: Aids mapping and read assignment between closely related genomes (e.g., Drosophila melanogaster vs Drosophila simulans).
• Low-quality or biased datasets: Suited for datasets with sequencing errors, evolutionary variation, or dataset-specific biases.

Methodology:

Implements probabilistic mapping using position-specific scoring matrices (PSSM), integrates sequencing read quality scores, prior mismatch probabilities, and user-defined evolutionary/bias models, computes mapping qualities, and is implemented as an extension of BWA retaining the BWA genome index structure.