GBScleanR

GBScleanR implements marker-specific error correction within a hidden Markov model to improve genotype calling from reduced-representation sequencing (RRS) data generated by next-generation sequencing (NGS).

Key Features:

• Marker-specific error rates: Integrates marker-specific error rates into the hidden Markov model to model heterogeneous error across markers.
• Allele read ratio bias modeling: Accounts for biases in allele read ratios arising from uneven amplification and read mismapping.
• HMM extension: Extends conventional hidden Markov model approaches to handle non-uniform error profiles across markers.
• RRS/NGS focus: Targets genotype data produced by reduced-representation sequencing platforms used in NGS-based genotyping.
• Empirical accuracy gains: Demonstrates substantial accuracy improvements in simulations (over 25 percentage points) and consistent reliable genotype estimation in real datasets.
• Robust handling of error-prone markers: Improves genotype estimation robustness for markers prone to errors.

Scientific Applications:

• High-throughput genotyping: Improves genotype quality control and error correction in high-throughput genotyping workflows using RRS.
• Population genetics: Enables more accurate genotype datasets for population genetics analyses.
• Evolutionary biology: Supports evolutionary biology studies that require reliable genotype calls.
• Plant breeding research: Enhances genotype reliability for plant breeding and related agricultural genetics studies.

Methodology:

Integrates marker-specific error rates into a hidden Markov model and models allele read ratio biases to perform genotype error correction on RRS-derived NGS data.