megasat: automated inference of microsatellite genotypes from sequence data

Area of Science:

• Genetics
• Bioinformatics
• Computational Biology

Background:

• Microsatellites are widely used genetic markers.
• Traditional microsatellite genotyping methods can be labor-intensive and prone to errors.
• Next-generation sequencing (NGS) offers high-throughput data generation capabilities.

Purpose of the Study:

• To introduce MegaSat, a novel software for automated microsatellite genotyping using NGS data.
• To evaluate the accuracy and efficiency of MegaSat in a real-world application.
• To demonstrate the potential of automated microsatellite analysis for large-scale genetic studies.

Main Methods:

• MegaSat processes pooled amplicon sequencing data to score microsatellite genotypes.
• The software incorporates fuzzy matching for sequencing errors and decision rules for amplification artifacts.
• Histograms of length-frequency distributions are generated for genotype evaluation, similar to electropherograms.

Main Results:

• MegaSat was used to genotype 1024 guppies at 43 microsatellites on an Illumina MiSeq.
• Mean genotyping error rates were estimated at 0.021 and 0.012 using pedigree and repeat data.
• High accuracy was achieved for most loci, with 26 loci showing error rates ≤0.007.

Conclusions:

• Automated microsatellite genotyping with MegaSat is feasible with high throughput and low error rates.
• Locus selection is crucial for optimizing accuracy in automated genotyping.
• MegaSat offers a cost-effective solution for large-scale microsatellite analysis in genetic research.