phred 33 from Ion S5?

Question

[Benn]

Hello,

I am using Ion S5 data from an amplicon panel, and with fastq_quality_trimmer I try to trim. However I get errors that my data contains characters which are not phred 33. Does anyone know why there are more characters in the quality score of these Ion S5 reads? And does anyone know how to get my trimming work?

My bash code is:

fastq_quality_trimmer -Q33 -t 30 -l 100 -i ${FQfile} -o ${FQfile%.fastq}.trim.fq

And the errors are:

fastq_quality_trimmer: Invalid quality score value (char 'K' ord 75 quality value 42) on line 24
fastq_quality_trimmer: Invalid quality score value (char 'J' ord 74 quality value 41) on line 60
fastq_quality_trimmer: Invalid quality score value (char 'J' ord 74 quality value 41) on line 40
fastq_quality_trimmer: Invalid quality score value (char 'J' ord 74 quality value 41) on line 4
fastq_quality_trimmer: Invalid quality score value (char 'K' ord 75 quality value 42) on line 8
fastq_quality_trimmer: Invalid quality score value (char 'L' ord 76 quality value 43) on line 12
fastq_quality_trimmer: Invalid quality score value (char 'L' ord 76 quality value 43) on line 8
fastq_quality_trimmer: Invalid quality score value (char 'M' ord 77 quality value 44) on line 152
fastq_quality_trimmer: Invalid quality score value (char 'J' ord 74 quality value 41) on line 108
fastq_quality_trimmer: Invalid quality score value (char 'J' ord 74 quality value 41) on line 80

Thanks in advance!

Ben

[GenoMax]

Technically the Sanger fastq scale does not need to stop at 40 (though that interval has been predominantly used).

It appears that S5 software is assigning Q-scores of >40 to some bases and fastx_toolkit must be coded to recognize values only to 40. If you want to keep using fastx then you may need to reduce Q-scores > 40 to 40 (or find a different trimming program).

[Benn]

Thanks for your explanation, can you suggest a tool for trimming?

[GenoMax]

You can use BBDuk from BBMap.
Keep in mind that BBDuk will floor (reduce the scores above 41 down to 41(=J) after trimming). You would want to use qtrim= and trimq= options based on your OP.