SiNPle: Fast and Sensitive Variant Calling for Deep Sequencing Data.

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• High-throughput sequencing generates vast amounts of genetic data, enabling high-resolution analysis.
• Deep sequencing is crucial for detecting rare variants in heterogeneous samples like viral quasi-species.
• Sequencing errors and PCR artifacts complicate accurate variant identification, especially at high read depths.

Purpose of the Study:

• To develop a fast and effective software tool for accurate variant calling from high-coverage sequencing data.
• To address the challenge of distinguishing true genetic variants from sequencing and PCR-induced noise.
• To provide a sensitive and specific solution for variant detection in complex biological samples.

Main Methods:

• Developed SiNPle (Simplified Inference of Novel Polymorphisms from Large coveragE), a software based on a simplified Bayesian approach.
• Incorporated individual base qualities, error rates (sequencing and PCR), and prior variant frequencies into the Bayesian model.
• Utilized an analytical formula for rapid computation of posterior probabilities, even with very high coverage data.

Main Results:

• SiNPle demonstrated superior speed and sensitivity compared to existing variant calling methods on simulated and real viral datasets.
• The software effectively filters putative single nucleotide polymorphisms (SNPs) and insertions/deletions (indels) based on user-defined sensitivity.
• The Bayesian model accurately assesses the probability of a variant being genuine versus an artifact.

Conclusions:

• SiNPle offers a computationally efficient and highly sensitive method for variant calling in deep sequencing data.
• The software is valuable for analyzing complex genetic samples, such as viral populations.
• SiNPle is freely available, promoting accessibility for researchers in bioinformatics and genomics.