Hello,
I am playing with Galaxy "splitters" capabilities. After some cases that you help me out to solve I am facing a new issue, this is maybe due to my tool configuration file, but in any case I tell you what I've done.
What I would like to do exactly, is to split paired fastq, map them and then join them. This is my configuration file:
<tool id="bwa_mio" name="map with bwa">
<description>map with bwa</description>
<parallelism method="basic" split_size="3" split_mode="number_of_parts" merge_outputs="output"></parallelism>
<command>
bwa mem -R '@RG\tID:foo\tSM:bar' /home/ralonso/BiB/Galaxy/data/Cclementina_v1.0_scaffolds.fa $input > temporary_bam_file.sam 2>/dev/null ;
samtools view -Sb temporary_bam_file.sam > temporary_bam_file.bam ;
samtools sort temporary_bam_file.bam $output ;
</command>
<inputs>
<param format="fastqsanger" name="input" type="data" label="fastq"/>
</inputs>
<outputs>
<data format="bam" name="output" />
</outputs>
<help>
bwa
</help>
</tool>
My problem of this configuration is that generates an empty file. So, after seeing the code, I discover that when it tries to join the several bam files it goes to the first parent: class Data( object ), to the method merge: def merge( split_files, output_file). So I may be wrong, but I think binary.bam class should override this method, is this right? if this is the case, I would like to implement this method, I have couple of basic ideas, like merge them with samtools. What do you think?
On ther other hand, is this related with the last email of John Chilton and the 15.03 release?
Best Regards
--
Roberto Alonso
Functional Genomics Unit
Bioinformatics and Genomics Department
Prince Felipe Research Center (CIPF)