Question 5 of 9 Advanced Concept
How do you identify contamination or sample swaps in sequencing data?
Short answer
Check sample identity and mixture signals separately. Compare genotype fingerprints with expected samples to detect swaps, then use allele patterns, contamination estimates, controls and batch metadata to investigate mixtures or cross-sample contamination. No single metric proves the cause.
Why
I would start with chain-of-custody records, sample IDs, barcodes, plate maps and negative controls, then inspect the sequencing data. For identity, compare high-confidence SNP fingerprints across library preparations or against a trusted sample from the same donor. Unexpected fingerprint mismatches can indicate a swap, whereas surprising matches may point to duplicates or mislabeled donors.
For contamination, inspect unexpected alternative alleles, allele-balance distributions, sex-chromosome signals where biologically appropriate, taxonomic content for relevant assays and assay-specific contamination estimates. Tools such as Picard CrosscheckFingerprints and GATK CalculateContamination address different situations and cannot be applied indiscriminately to every platform or specimen type. Check whether the anomaly is isolated to one sample, a plate, an extraction batch or a sequencing lane.
Escalate suspected identity failures and refrain from making clinical or scientific claims until identity and data integrity are resolved.
Common mistakes
- Assuming low mapping rate always means contamination, or interpreting a tool designed for somatic samples as universally valid for all sequencing assays
The interviewer might ask next
How would you distinguish an accidental plate swap from low-level cross-contamination?