Sequencing Depth & Data Volume Calculator (Lander-Waterman)

Sequencing depth (coverage / depth) is defined by the Lander-Waterman model:

C = N × L / G

• C: mean depth, expressed as 30×, 100×, etc.
• N: number of reads
• L: read length (for paired-end sequencing, use the combined length of both strands; PE150 → 300 bp)
• G: genome size (or target region size)

Conversely, the data volume needed to reach a target depth = C × G; divide by read length to get read count.

Model assumptions, and when they break down:

Lander-Waterman assumes reads are uniformly and randomly distributed across the genome. In practice this assumption often fails:

• GC bias: regions with very high or very low GC content tend to have significantly lower coverage;
• Repetitive sequences: reads from multi-copy regions cannot be uniquely mapped, so effective depth falls below the mean;
• Capture-based sequencing: whole-exome and panel assays typically have an on-target rate of only 60–80%. This discount must be factored into data volume calculations, or actual depth will fall well short of expectations.

The value computed here is therefore the theoretical mean depth. True uniformity requires examining the coverage distribution after alignment, not just this average.

Note that losses from low-quality filtering depend on the quality distribution of the data itself — to decode the quality string in a FASTQ file, use the FASTQ Quality Score Converter. When tallying covered intervals after alignment, watch for the 1-based vs. 0-based coordinate offset between BED and GFF — see the Genome Coordinate System Converter.