[QUESTION] Guessing FASTQ quality system using seqkit convert

[alienzj]

seqkit version

➤ seqkit version
seqkit v0.16.1

Problem I meet:

➤ seqkit convert V300109217_L03-506_1.fq.gz | head
[INFO] possible quality encodings: []
[ERRO] quality encoding not consistent

Dear Shenwei,

I used seqkit to guess the FASTQ quality system, but the quality system does not seem to be identified.
Any good suggestions ?

Thank you for developing seqkit and excellent documentation.

[alienzj]

➤ seqkit version
seqkit v2.0.0

➤ seqkit convert V300109217_L03-506_1.fq.gz | head
[INFO] possible quality encodings: []
[ERRO] quality encoding not consistent

[shenwei356]

What's the sequencing platform?

[alienzj]

What's the sequencing platform?

Transcriptome Resequencing based on DNBseq platform.

Below is FastQC report:

[shenwei356]

will check it on Friday.

[shenwei356]

Similar issue #239 .

I've extended the value range of 'QualityEncoding', which should correct this.

• seqkit_windows_amd64.exe.tar.gz
• seqkit_darwin_amd64.tar.gz
• seqkit_linux_amd64.tar.gz

[alienzj]

Hi, Shenwei, thanks for your help.

I downloaded the latest compiled version of seqkit you provided: seqkit_linux_amd64.tar.gz,
but the quality system still does not seem to be identified.

➤ ./seqkit version
seqkit v2.0.0

➤ ./seqkit convert V300109217_L03-506_1.fq.gz | head
[INFO] possible quality encodings: []
[ERRO] quality encoding not consistent

[shenwei356]

paste some sequences please.

[alienzj]

test_100K.fq.gz

Dear Shenwei, thanks,
please see the attachments for 100K reads which were randomly generated from a bigger data sets.

You will see the following result:

➤ ./seqkit convert test_100K.fq.gz | head
[INFO] possible quality encodings: []
[ERRO] quality encoding not consistent

[shenwei356]

Some reads are treated as Illumina 1.5. It can be solved by using a small value of -N/--thresh-B-in-n-most-common, like 2.

I also changed the default value from 4 to 2, related discussion.

• seqkit_linux_amd64.tar.gz

[alienzj]

➤ ./seqkit convert test_100K.fq.gz | head
[INFO] possible quality encodings: [Sanger Illumina-1.8+]
[INFO] guessed quality encoding: Sanger
[INFO] converting Sanger -> Sanger
[WARN] source and target quality encoding match.

Great, Shenwei, thank you sou much !