MarkerMatch: A Proximity-Based Probe-Matching Algorithm for Joint Analysis of Copy-Number Variants from Different...

Area of Science:

• Genetics
• Bioinformatics
• Genomic analysis

Background:

• Copy-number variants (CNVs) are crucial in complex human disorders.
• Current CNV detection methods using microarrays are limited by probe overlap across array types, reducing sensitivity.
• Existing approaches often use a consensus probe set, which can excessively decrease overall sensitivity.

Purpose of the Study:

• To develop and validate MarkerMatch, a proximity-based algorithm for CNV calling.
• To overcome limitations of differing array-specific sensitivities in CNV detection.
• To enhance the resolution and sensitivity of CNV calling while maintaining precision.

Main Methods:

• Developed MarkerMatch, a proximity-based algorithm to match probes across different genotyping microarrays.
• Applied MarkerMatch to CNV calls from 4,906 individuals genotyped on three array types (Global Screening Array, Omni2.5, Omni Express Exome).
• Optimized MarkerMatch parameters (D_MAX and Method) for CNV calling.

Main Results:

• MarkerMatch significantly improves CNV detection sensitivity by increasing probe density.
• The algorithm maintains or improves precision compared to current practices, such as using only consensus probes.
• MarkerMatch outperforms current methods in F1 score, Fowlkes-Mallows index, and Jaccard index.
• An optimal D_MAX setting was identified at 10kb, with Method parameter suitability dependent on the specific use case.

Conclusions:

• MarkerMatch offers a superior approach for CNV calling across diverse microarray platforms.
• The algorithm enhances the ability to detect CNVs, crucial for understanding complex human diseases.
• MarkerMatch provides a more sensitive and precise method for genomic association studies involving CNVs.