FASTQ Quality Viewer

Quality Control Guidelines

How FASTQ Quality Scores Work

A FASTQ record has four lines: a header beginning with @, the nucleotide sequence, a separator line beginning with +, and a line of quality characters — one per base. Each quality character encodes a Phred quality score (Q), which expresses the sequencer's confidence that the base was called correctly. The score is defined on a logarithmic scale as Q = -10 × log10(P), where P is the estimated probability that the base call is wrong. A higher Q means a lower error probability.

Because of the logarithmic scale, Q20 corresponds to a 1-in-100 chance of error (99% accuracy) and Q30 to a 1-in-1,000 chance (99.9% accuracy). The proportion of bases at or above these thresholds, often reported as %Q20 and %Q30, is a standard yardstick for run quality in next-generation sequencing. Quality characters are stored as printable ASCII symbols: the Phred+33 (Sanger and Illumina 1.8+) encoding adds 33 to the score to pick a character, while the older Phred+64 encoding adds 64. This viewer inspects the character range in your data to detect which offset applies before converting symbols back into numeric scores.

Common Use Cases

• Checking raw reads straight off a sequencer before committing to a full analysis
• Deciding where to trim low-quality bases, which often accumulate toward the 3' end of reads
• Comparing %Q30 across samples or lanes to spot a failing run early
• Confirming quality encoding when combining data from different platforms or archives
• Teaching students how Phred scores and the FASTQ format relate to base-call accuracy

Frequently Asked Questions

What do Q20 and Q30 actually mean?

Q20 means the sequencer estimates a 1-in-100 probability that the base is wrong, equivalent to 99% accuracy. Q30 means a 1-in-1,000 probability, or 99.9% accuracy. Reporting the percentage of bases at or above Q30 is a widely used summary of overall run quality.

What is the difference between Phred+33 and Phred+64 encoding?

Both encodings store quality scores as ASCII characters, but they use different offsets. Phred+33 adds 33 to the numeric score to choose the character and is used by Sanger and modern Illumina data. Phred+64 adds 64 and appears in older Illumina 1.3 to 1.7 files. Using the wrong offset shifts every score by 31, so detecting the encoding correctly is essential.

Why does read quality drop off toward the end of the read?

In sequencing-by-synthesis, signal quality tends to decline over successive cycles as the chemistry becomes less synchronized across a cluster of molecules. As a result, bases near the 3' end of a read are frequently lower quality, which is why per-position plots and quality trimming are so useful.

Should I trim or filter low-quality bases?

Often, yes. Trimming low-quality tails and removing adapter contamination before alignment or assembly improves the reliability of downstream results. The right thresholds depend on your application, so use the per-position and summary statistics here to decide how aggressive to be.

What does an N in the sequence line mean?

An N indicates a position where the sequencer could not confidently call a specific base. A small number of Ns is normal, but a high proportion — especially clustered at particular positions — can signal a problem with the run or the sample.