fastp FILTER results??

[fatyang799]

The relationship between character and Q score:

• !=05
• 4=19
• 5=20
• 6=21
• ?=30

To understand the mechanism of fastp trimming reads, I design a fastq file: test.fastq

the @line indicate the median and mean Q value of the read:

@median=17.5;mean=17.5
GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG
+
??????????????????!!!!!!!!!!!!!!!!!!
@median=30;mean=18.89
GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG
+
????????????????????!!!!!!!!!!!!!!!!
@median=20;mean=20.17
GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG
+
666666666666655555555555555554444444
@median=20;mean=19.81
GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG
+
555555555555555554444444444444666666

Then I run the code:

fastp -i test.fastq -q 10 -o out.fastq

The results:

@median=20;mean=20.17
GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG
+
666666666666655555555555555554444444
@median=20;mean=19.81
GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG
+
555555555555555554444444444444666666

It was weird! And I do not know how to read C++, so I really do not understand why it turns out like this.

I set the cutoff of Q value should be 10. And I thought fastp would trim bases with low quality (like: !), but it just delete all read install of bases with low quality.