isoseq-collapse-on-merged-bam: A Shell repository from ylipacbio

There is a bug in SMRTLink 7.0 IsoSeq with Mapping workflow, which was 
- introduced by chunking transcripts bam file into multiple small files, 
- while IsoSeq collapsing is performed on each of the chunk (each small file),
thus,
- transcripts mapping to the same genomic location (which should be placed in the same group), may be put into different gene groups. 

Please wait for official bug fix bug in SMRTLink 8.0 release!

Note this is not official bug fix, this is NOT official hot patch. 
This is NOT ISO-compliant and there is No guarantee, no libility.
For Advanced Users willing to try at their own risk, here it is.

Prerequisite:
  You have access to SMRTLink 7.0 build for command line tools such as
  `dataset`, `python -m isocollapse.tasks.collapse_mapped_isoforms`, `samtools`, `python`

Steps:
1. Download `isocollapse-on-merged-bam.sh` and `reset-transcriptalignmentset.py`

  wget https://github.com/ylipacbio/isoseq-collapse-on-merged-bam/blob/master/isocollapse-on-merged-bam.sh
  wget https://github.com/ylipacbio/isoseq-collapse-on-merged-bam/blob/master/reset-transcriptalignmentset.py

2. identify job root of SMRTLink IsoSeq with Mapping job, ${job_root}, and 
3. then call 

   bash isocollapse-on-merged-bam.sh  ${job_root}
   
e.g., if SMRTLink IsoSeq job path is: /mymount/myjob/001/0012345, run

   bash isocollapse-on-merged-bam.sh  /mymount/myjob/001/0012345

You will see regenerated output files out.gff, out.fastq etc