How much sequencing coverage do I need?
Coverage is the average number of times each base in your target is read. Choosing it well is the difference between confident calls and wasted reads.
What coverage means
Mean depth of coverage is total sequenced bases divided by the size of what you are sequencing: depth = (read length × read count) ÷ genome size. A 30× genome has, on average, 30 reads spanning every position — but it is an average, so some regions sit well below it.
Typical targets by application
- Human whole-genome (WGS), germline: ~30× is the common standard for confident SNV and small-indel calling.
- Whole-exome (WES): ~100× mean on target, because capture is uneven and you want most bases comfortably covered.
- Somatic / cancer: 100× and up for the tumour (often 200–500× for low-frequency variants), with a matched normal at ~30×.
- Bacterial / small genomes: 30–100× is plentiful for assembly and variant calling.
- RNA-seq: measured in reads per sample (e.g. 20–30M reads for differential expression), not genome depth, because transcript abundance varies enormously.
Planning a run
Work backwards from the depth you need to the number of reads to order, and remember the headline figure is a theoretical mean: duplicates, off-target reads and quality trimming all reduce usable depth, so build in margin.