how to adjust phred33 and phred64 in a fastq

Question

[zhangmingpeng]

I downed a fastq file from NCBI whose quality code was mixed with phred33 and phred64. In the process of mapping, the program will be mistaken. General software does not correct this mixed situation, so I want to write a useful code, but I have limited ability, so I need help.

[WouterDeCoster]

How did you figure out that one file contains both quality encodings? How does that even happen?

[zhangmingpeng]

When I used BWA for mapping without '-fixMisencodedQuals', the error message showed " we encountered an extremely high quality score of 64"; while when I add '-fixMisencodedQuals', the error message showed "while fixing mis-encoded base qualities we encountered a read that was correctly encoded". So I guess this file contained both quality encodings.

[zhangmingpeng]

But I don't know how this happened.

Answer 1

[chen]

phred64 is out of date, and it's recommended to convert phred64 FASTQ to phred33 FASTQ before downstream analysis. You can google it to find a lot of tools for this conversion.

[zhangmingpeng]

yeah, I have tired some tools, such as 'seqkit', but they can not successfully convert the the part of phred64 FASTQ to phred33 FASTQ in such mixed FASTQ. So someone advised me to write a code myself to deal with it. I need help!