Joint variant and de novo mutation identification on pedigrees from high-throughput sequencing data

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Accurate identification of genetic variants from next-generation sequencing (NGS) data is crucial for genetic disease studies.
• Existing methods often overlook familial relationships, leading to substantial Mendelian errors and reduced accuracy.
• Analyzing trio and pedigree data is essential for identifying disease-causing mutations.

Purpose of the Study:

• To develop a novel Bayesian network framework for joint analysis of pedigree sequencing data.
• To improve the accuracy and reliability of genetic variant identification and genotyping in family studies.
• To enable scalable analysis of large pedigrees and detect de novo mutations.

Main Methods:

• Developed a Bayesian network framework integrating Mendelian segregation priors.
• Jointly analyzed whole-genome sequencing (WGS) data from all members of a pedigree simultaneously.
• Evaluated the method using simulations and WGS data from a 17-individual, 3-generation CEPH pedigree.

Main Results:

• The family caller significantly improved the quality of identified variants compared to singleton calling.
• Eliminated spurious calls, as evidenced by improved Ti/Tv ratios, Het/Hom ratios, and concordance with dbSNP/SNP array data.
• Successfully identified all validated de novo mutations in NA12878 with a 7× precision improvement.

Conclusions:

• The proposed Bayesian network framework offers a scalable and accurate solution for analyzing pedigree sequencing data.
• This method enhances the identification of genetic variants and de novo mutations in large-scale genomics and human disease studies.
• Integrating familial relationships improves the robustness of variant calling in genetic analyses.