Is the PRINSEQ fastq-format weird?

[JonB]

Hi,

I have been using Prinseq to trim my Illumina data, and the fastq output looks like this:

@HWI-ST486:386:D1UMHACXX:3:1101:1399:2119 1:N:0:TGACCA
TCGTATCTGTAATCATGAACTTGTCAACGGCTACCTGGTTTCTGTCCT
+HWI-ST486:386:D1UMHACXX:3:1101:1399:2119 1:N:0:TGACCA
1:BDBEDFHHGHGCBGHGCHHIJGIFGGIGEHCHJJIJHGH<FDGDHF
@HWI-ST486:386:D1UMHACXX:3:1101:1367:2139 1:N:0:TGACCA
ATGTTTTTTGGGGTTATAACAGGGTGGAGCGCTTTATGCGACTTCGCCCTTT
+HWI-ST486:386:D1UMHACXX:3:1101:1367:2139 1:N:0:TGACCA
1=DDDEDHHDHHJJIGIFDFGIIJHGI?GHJIFHI@GG@F@;AH=?=BDCEE

Why is the header repeated twice? Mapping with bowtie seems to be fine, but when I try to collapse identical reads with fastx it will not take the format.

[mastal]

Is the PRINSEQ fastq-format weird?

The specification for fastq is that the third line starts with +, optionally followed by the read identifier.

[JonB]

I want to use FastX to collapse identical reads in the Fastq file and keep the count in the header.

It seems that Prinseq only outputs fastq with the identifier in the third line, and that fastx will not take this input.

Do you know of other programs that will collapse in this way?

[ddb]

Prinseq option -no_qual_header should only show + in the third line.

[yzzhang]

fastx? you mean the tool inside fastx toolkit? have you tried to define the quality score parameter -Q 33?

[SES]

Prinseq option -no_qual_header should only show + in the third line.

That is correct, but I can see where it could be confusing for some people that the default for prinseq is to print the qual header when most parsers default to not printing this header. Previous versions of the program did not have this option, which was annoying because you had to run another command to remove this line.

[JonB]

Thanks for the replies! I hadn't noticed that option in Prinseq

[drdna]

Anyone have any idea what prinseq's 5maxd, 3maxd and exactmaxd stats mean? There is no explanation in the manual.

[ddb]

It is related to duplicates. This section from the help menu has some explanation.

-stats_dupl
Outputs the number of exact duplicates (exact), 5' duplicates
(5), 3' duplicates (3), exact duplicates with reverse
complements (exactrevcom) and 5'/3' duplicates with reverse
complements (revcomp), and total number of duplicates (total).
The maximum number of duplicates is given under the value name
with an additional "maxd" (e.g. exactmaxd or 5maxd).