Bioinformatics and Data Interpretation
Core concepts
Understand the NGS pipeline: library prep, sequencing, base calling, alignment to a reference genome, variant calling, annotation, and filtering. Know coverage/depth, read quality (Q scores), mapping quality, variant allele fraction (VAF), and how these affect confidence. Be familiar with reference genome builds (GRCh37/hg19 vs GRCh38/hg38) and the importance of consistent build reporting.
Know the variant filtering workflow: frequency filters (gnomAD), impact prediction, inheritance-based filtering (trio analysis), and phenotype-driven prioritization (HPO terms). Understand copy-number calling from NGS data, structural variant detection, and the role of orthogonal confirmation. Recognize batch effects, contamination, and sample swap signals.
Traps
Common traps: interpreting a variant at low VAF without considering mosaicism vs sequencing error; ignoring reference build mismatch; and treating a VUS as actionable. Another trap: assuming high depth guarantees accuracy (systematic errors and mapping artifacts persist). Watch for trio analysis misinterpretation - de novo calls require confirmation.
Rule of thumb
A variant is only as good as its coverage and its confirmation. Check depth, VAF, mapping quality, and orthogonal confirmation before reporting a clinically significant call.
Worked example
A de novo candidate variant in a proband–parent trio shows 45% VAF in the proband and 0% in parents, but coverage is 8x. Low depth undermines confidence; the variant should be confirmed by Sanger sequencing before reporting as de novo.