Does not correctly decode old Illumina Phred score encoding

[shigita]

I ran Falco 0.2.4 and FastQC on old Illumina data, and noticed that the recognition of Phred score encoding differs between programs as follow:
Falco recognized as Illumina 1.9, and Per base sequence quality and Per seqence quality score look incorrect.
FastQC recognized as Illumina 1.5, and everything looks fine.

Attached, small example of old fastq and output from each program.

• read_1.fastq.gz
• falco_result.zip
• fastqc_result.zip

Here are the commands I ran:

# Falco
falco --nogroup read_1.fastq.gz --outdir falco_result

# FastQC
fastqc --nogroup read_1.fastq.gz --outdir fastqc_result

I hope these help reproduce the problem.
Thanks in advance.

[guilhermesena1]

Hello,

Thank you for providing the data to reproduce the issue! We have some issues open regarding quality distributions on both nanopore and legacy phred32 from Illumina. Currently the program assumes everything is Illumina, and we will address this problem in the upcoming release. I sincerely apologize for the inconvenience!

[guilhermesena1]

with your data I was able to push a fix for the problem at baad210 . I haven't tested it thoroughly, but for your dataset at least it now gives the same plots for per base sequence quality, basic statistics and per sequence quality scores. Thank you once again for communicating the issue!

[shigita]

I have confirmed that it works even on my whole data!
Thank you very much for your prompt response 😊

[guilhermesena1]

glad to hear it! I'll close this for now just to keep track of current issues that need addressing but feel free to reopen if you run into any problems with the current push.