Open ghuls opened 4 years ago
Thanks for reporting! I am currently re-writing tagdust to make it more robust and to include read/writing to compressed files.
tagdust
seems like a great tool. I was looking for a program like this for many years. To to something similar with cutadapt
seems to be quite difficult (although they should have most of the basic building block implemented).
For this specific project I wanted to extract a UMI-like 17bp sequence from the FASTQ file. when using -3 R:NNNNNNNNNNNNNNNNN
the extracted reads are not all 17bp (rest of the read is useless). It would be nice if it would be possible to specify the exact read length is this case (and ignore reads that didn't match). I easily can do a simple post filter to fix it manually now.
When using -3 F:NNNNNNNNNNNNNNNNN
with no R:
specified, tagdust
segfaults.
$ tagdust -t 8 -e 0.1 -1 O:N -2 S:ACTCGAGCCGAGCAGGCGCGCCGATC -3 F:NNNNNNNNNNNNNNNNN -4 S:GGACCGGG -5 O:N -o test test.fastq.gz
[2019-12-20 11:47:13] Tagdust 2.32, Copyright (C) 2013-2019 Timo Lassmann <timolassmann@gmail.com>
[2019-12-20 11:47:13] cmd: tagdust/src/tagdust -t 8 -e 0.1 -1 O:N -2 S:ACTCGAGCCGAGCAGGCGCGCCGATC -3 F:NNNNNNNNNNNNNNNNN -4 S:GGACCGGG -5 O:N -o test test.fastq.gz
[2019-12-20 11:47:13] Start Run
--------------------------------------------------
[2019-12-20 11:47:14] Determining threshold for read0.
[2019-12-20 11:47:18] Long sequence found. Need to realloc model...
[2019-12-20 11:48:07] Selected Threshold:: 0.013495
4
Fatal exception (source file io.c, line 853):
malloc of size 0 failed
Aborted (core dumped)
tagdust segfaults when output files exists already.
It crashes in the
free_read_structure(param->read_structure)
call forread_structure->numseq_in_segment[0]
:"Debugging" code: