ExpressionAnalysis / STAR-SEQR

RNA Fusion Detection and Quantification
Other
16 stars 12 forks source link

Error: no SAMPLE.Chimeric.out.junction file #18

Open personalis opened 6 years ago

personalis commented 6 years ago

When trying to run STAR-SEQR, I get this error after STAR has finished alignment: OSError: [Errno 2] No such file or directory: '/path/to/SAMPLE.Chimeric.out.junction'

The reported STAR command line specifies "--chimOutType SeparateSAMold". Reading the STAR manual, in order to get the SAMPLE.Chimeric.out.junction file, I need to specify "--chimOutType Junctions". Is that correct, or should I expect the .junction file to be part of the output from the STAR command as generated by STAR-SEQR?

EDIT: my output folder DOES contain a file named SAMPLE.SJ.out.tab. Is this perhaps the expected junctions file by a different name? The content looks like this: 1 12228 12612 1 1 1 0 1 44 1 12698 13482 1 1 0 1 0 58 1 12722 13220 1 1 1 0 1 64 1 14830 14969 2 2 1 59 361 72

EDIT2: the SAMPLE.SJ.out.tab file is not the same as SAMPLE.Chimeric.out.junction. I tried adding a symlink in the code, but then I get a parsing error due to missing columns in the file.

EDIT3: I tried changing --chimOutType to Junctions, but then i get an error that the SAM file doesn't exist. So I definitely need some way to have STAR output both the SAM and the junctions file. I tried setting --chimOutType to both values (space- and comma-delimited), but STAR wouldn't accept that.

personalis commented 6 years ago

Update: I figured out how to send both "Junctions" and "SeparateSAMold" as arguments to the --chimOutType parameter. STAR-SEQR is running now, in the STAR alignment stage.

My patch is very simple, I just added 'Junctions' to this part of star_funcs.py, in the run_star() function: ` STAR_args = ['STAR', '--readFilesIn', fq1, fq2, '--readFilesCommand', 'zcat',
'--runThreadN', str(threads), '--genomeDir', star_index,
'--outFileNamePrefix ', prefix + ".", '--chimScoreJunctionNonGTAG', -1,
'--outSAMtype', 'None', '--chimOutType', 'Junctions', 'SeparateSAMold',

'--outSAMtype', 'BAM', 'SortedByCoordinate', '--chimOutType', 'WithinBAM',

                 '--alignSJDBoverhangMin', 5, '--outFilterMultimapScoreRange', 1,                                                                                                                                                                                                                                       
                 '--outFilterMultimapNmax', 5,                                                                                                                                                                                                                                                                          
                 '--outMultimapperOrder', 'Random', '--outSAMattributes', 'NH', 'HI', 'AS', 'nM', 'ch']                                                                                                                                                                                                                 

`

jasper1918 commented 6 years ago

Hi @personalis - Thanks for reporting the issue. Can you let me know which version of STAR you are using?

It looks like STAR changed the behavior of --chimOutType on the 2.6 releases. I'll have to add some requirements around versions or handle the differences within the code going forward.

nservant commented 4 years ago

Hi, Any stable patch for this issue please ? It makes STAR-seqr uncompatible with the most recent version of STAR. Thanks