Ancestry-agnostic estimation of DNA sample contamination from sequence reads

Area of Science:

• Genomics
• Population Genetics
• Bioinformatics

Background:

• DNA sample contamination impacts genotype quality and downstream analysis.
• Existing methods require accurate population allele frequencies, which are often impractical to specify for diverse, large-scale studies.
• Misspecified allele frequencies can lead to biased contamination rate estimations and failure to identify contaminated samples.

Purpose of the Study:

• To develop a robust statistical method for accurate DNA contamination estimation.
• To create a method that is agnostic to the genetic ancestry of the intended or contaminating sample.
• To simultaneously estimate genetic ancestry and DNA contamination, correcting for biases in existing methods.

Main Methods:

• A unified likelihood framework integrating genetic ancestry and DNA contamination estimation.
• Leveraging individual-specific allele frequencies projected from reference genotypes onto principal component coordinates.
• A novel statistical approach for robustly estimating contamination and ancestry.

Main Results:

• The proposed method accurately estimates DNA contamination rates across diverse populations and contamination scenarios.
• The method successfully estimates genetic ancestries while accounting for potential contamination.
• Demonstrated robustness in contamination and ancestry estimation compared to existing methods.
• Showed that ignoring contamination can substantially bias genetic ancestry inference, a bias corrected by the new method.

Conclusions:

• The developed method provides a robust solution for DNA contamination detection and genetic ancestry estimation.
• This approach enhances the reliability of genotype calls and downstream analyses in large-scale sequencing projects.
• The method offers improved accuracy and corrects biases inherent in current contamination and ancestry inference techniques.