Error! Quality score file missing values for score range

[kingdynasty]

My fastq file is Illumina-1.8 Phred+33 format, how to solve this problem?

[jsh58]

As described here, the quality score profiles of NGmerge are limited to the range [0, 40]. If you have scores outside that range, you would need to provide alternative profiles. This can be accomplished by modifying qual_profile.txt, adding rows and columns to each matrix.

[mchimenti]

kingdynasty, I had this same issue. Let me give a clearer answer:

You need to use Phred+33 (0,41). That means setting '-u 41' instead of the default (40). This will allow "J" to be an acceptable quality score character.

[tiantaibird]

As described here, the quality score profiles of NGmerge are limited to the range [0, 40]. If you have scores outside that range, you would need to provide alternative profiles. This can be accomplished by modifying qual_profile.txt, adding rows and columns to each matrix.

I have the same issues. What numbers shoul I add for the rows and columns to each matrix?

[marcgall]

I have the same issue, and it is not solved by setting '-u 41' by itself: The -w parameter must also be supplied with a modified qual_profile.txt. Modifying the matrixes with an extra row and column worked (I simply copied the outermost values). But isn't it odd that this problem crops up for such a common quality scheme?

[jsh58]

Not odd at all; the matrices were created by analyzing sequence data that was limited to [0, 40] (cf. Figure 2).

[czhu]

can you provide an option for quality score 41 since this is so common nowadays? I get the reason but it's quite common now to have Illumina 1.9 quality score, this issue really prevent us from using this nice tool.

[damienmarsic]

I tried NGmerge for the first time today and I have the same problem. the option '-u 41' does not help. Adding a row and a column to qual_profile.txt does not help either, unless I did not do it correctly. It would be nice to have a valid qual_profile.txt file to download.

[Jdbeck66]

Same issue. Seems like some explanation of how a custom profile is constructed would be a useful addition to the documentation. Suggesting that adding rows and columns to the matrix probably isn't sufficiently detailed to be helpful and pointing to a chart as the basis for the default profile doesn't indicate how that data was used to create the profile. This tool is a really useful tool for getting the relative abundances correctly identified in my downstream counting algorithms - so I'm hopeful you'll see this as a useful suggestion.

[jsh58]

The format of a file with custom quality score profiles is described in the README under -w <file>. The process of creating the default profiles is thoroughly described in the NGmerge manuscript.

[damienmarsic]

Easy fix: add both -u 41 and -g arguments.
Example:
NGmerge -1 Reads_1.fq.gz -2 Reads_2.fq.gz -o Merged_reads.fq.gz -u 41 -g