Reads, read length and sequencing yield

A sequencer’s spec sheet quotes an output in gigabases (Gb) or a number of reads. Converting between the two is simple arithmetic — once you keep track of read length and whether the run is paired-end.

The formula

yield (Gb) = reads × read length × mates ÷ 1,000,000,000

Mates is 2 for paired-end and 1 for single-end. A paired-end run of 400 million read pairs at 2 × 150 bp: 400,000,000 × 150 × 2 = 120,000,000,000 bp =120 Gb. Each pair contributes both a forward and a reverse read, so the two mates count together.

Going the other way

To size a run to a target yield, rearrange:

reads = target Gb × 10⁹ ÷ (read length × mates)

For 90 Gb of PE150: 90×10⁹ ÷ (150 × 2) = 300 million read pairs. This is the figure to check a platform’s read-count spec against when you know how much data you need.

Reads vs read pairs

The common source of confusion is whether a number counts reads or pairs. Paired-end specs almost always quote pairs (or “clusters”), and each pair is two reads. If a tool asks for reads and you have pairs, keep the layout set to paired-end so both mates are counted — don’t double the number yourself.

Raw is not usable

Yield is raw output. Duplicate reads, adapter and quality trimming, and reads that don’t map all shave it down before it becomes usable coverage. Budget headroom above the raw Gb figure, then pair this with a coverage calculation to turn gigabases into depth on your genome.