What is Q30? Phred quality scores explained
A Phred quality score (Q) is a sequencer's confidence in a single base call, on a log scale. Q30 — 99.9% accuracy — is the everyday benchmark for short-read data.
The Phred scale
The score relates to the probability P that a base call is wrong: Q = −10 × log₁₀(P), so P = 10−Q/10 and accuracy is 1 − P. Because it is logarithmic, every 10 points multiplies the accuracy by ten.
- Q10 → P = 0.1 → 90% accurate (1 error in 10).
- Q20 → P = 0.01 → 99% accurate (1 error in 100).
- Q30 → P = 0.001 → 99.9% accurate (1 error in 1,000).
- Q40 → P = 0.0001 → 99.99% accurate (1 error in 10,000).
Why %≥Q30 matters
Run reports quote the percentage of bases at or above Q30 as a headline quality metric. A high %≥Q30 means most base calls are trustworthy; a score that drops toward the 3′ end of reads is the usual reason to quality-trim before alignment or assembly.
Encoding
In a FASTQ file each base's quality is a single character. Modern Illumina data uses Phred+33: subtract 33 from the character's ASCII code to get Q. So "I" (ASCII 73) is Q40, and "!" (ASCII 33) is Q0.