Using strandness flag results in 0 LSVs, using strandness=none gives LSVs

[Anna-Leigh Brown]

Hi there,

I've had an unusal case with my data where using the strandness flag results in 0 LSV but using stranded none gives results....

The data is single-end fr_secondstrand;

Using either forward or reverse flag I get the same results:

2019-08-28 12:09:11,802 (PID:400283) - INFO - Majiq Build v2.1-c3da3ce
2019-08-28 12:09:11,803 (PID:400283) - INFO - Command: /annotation/mouse/gencode/gencode.vM22.annotation.gff3 -c wt_reverse_test.tsv -j 16 -o /majiq/builder_simple_rev/ --simplify
2019-08-28 12:09:11,803 (PID:400283) - INFO - Parsing GFF3
2019-08-28 12:10:04,151 (PID:400283) - INFO - Reading bamfiles
2019-08-28 12:10:04,159 (PID:400283) - INFO - Group wt_base, number of experiments: 3, minexperiments: 2
2019-08-28 12:10:04,159 (PID:400283) - INFO - Reading bam file WT1A.Aligned.sorted.out
2019-08-28 12:12:12,627 (PID:400283) - INFO - Detect Intron retention WT1A.Aligned.sorted.out
2019-08-28 12:12:33,050 (PID:400283) - INFO - Done Reading file WT1A.Aligned.sorted.out
2019-08-28 12:12:33,548 (PID:400283) - INFO - Reading bam file WT2A.Aligned.sorted.out
2019-08-28 12:14:07,823 (PID:400283) - INFO - Detect Intron retention WT2A.Aligned.sorted.out
2019-08-28 12:14:27,819 (PID:400283) - INFO - Done Reading file WT2A.Aligned.sorted.out
2019-08-28 12:14:28,371 (PID:400283) - INFO - Reading bam file WT3A.Aligned.sorted.out
2019-08-28 12:16:05,149 (PID:400283) - INFO - Detect Intron retention WT3A.Aligned.sorted.out
2019-08-28 12:16:25,089 (PID:400283) - INFO - Done Reading file WT3A.Aligned.sorted.out
2019-08-28 12:16:25,705 (PID:400283) - INFO - Detecting LSVs ngenes: 55487
2019-08-28 12:16:27,006 (PID:400283) - INFO - Starting simplification True
2019-08-28 12:16:41,574 (PID:400283) - INFO - Finished simplification
2019-08-28 12:17:55,232 (PID:400283) - INFO - 0 LSV found
2019-08-28 12:17:55,234 (PID:400283) - INFO - DUMP file b'WT1A.Aligned.sorted.out'
2019-08-28 12:18:26,246 (PID:400283) - INFO - Create majiq file
2019-08-28 12:18:26,311 (PID:400283) - INFO - Dump majiq file
2019-08-28 12:18:26,326 (PID:400283) - INFO - WT1A.Aligned.sorted.out: 0 LSVs
2019-08-28 12:18:26,327 (PID:400283) - INFO - DUMP file b'WT2A.Aligned.sorted.out'
2019-08-28 12:18:38,673 (PID:400283) - INFO - Create majiq file
2019-08-28 12:18:38,723 (PID:400283) - INFO - Dump majiq file
2019-08-28 12:18:38,725 (PID:400283) - INFO - WT2A.Aligned.sorted.out: 0 LSVs
2019-08-28 12:18:38,726 (PID:400283) - INFO - DUMP file b'WT3A.Aligned.sorted.out'
2019-08-28 12:19:03,553 (PID:400283) - INFO - Create majiq file
2019-08-28 12:19:03,605 (PID:400283) - INFO - Dump majiq file
2019-08-28 12:19:03,609 (PID:400283) - INFO - WT3A.Aligned.sorted.out: 0 LSVs
2019-08-28 12:19:07,063 (PID:400283) - INFO - MAJIQ Builder is ended successfully!

However with strandness=none:

2019-08-28 11:56:47,588 (PID:344147) - INFO - WT1A.Aligned.sorted.out: 38772 LSVs
2019-08-28 11:56:47,627 (PID:344147) - INFO - DUMP file b'WT2A.Aligned.sorted.out'
2019-08-28 11:57:17,947 (PID:344147) - INFO - Create majiq file
2019-08-28 11:57:20,588 (PID:344147) - INFO - Dump majiq file
2019-08-28 11:57:21,017 (PID:344147) - INFO - WT2A.Aligned.sorted.out: 38772 LSVs
2019-08-28 11:57:21,048 (PID:344147) - INFO - DUMP file b'WT3A.Aligned.sorted.out'
2019-08-28 11:57:51,857 (PID:344147) - INFO - Create majiq file
2019-08-28 11:57:54,193 (PID:344147) - INFO - Dump majiq file
2019-08-28 11:57:54,676 (PID:344147) - INFO - WT3A.Aligned.sorted.out: 38772 LSVs
2019-08-28 11:57:54,714 (PID:344147) - INFO - DUMP file b'WT1B.Aligned.sorted.out'

[salim megat]

Hi, 

Did you figure out where the problem was ? I am facing the same issue with my data being stranded but not having results after simplification....

many thanks, 

Salim. 

[Anna-Leigh Brown]

Hi there Salim,

No I could not figure out what was causing this error.

Have left it on strandedness none for now.

Would appreciate if the authors have any advice why this could be happening. At first I thought it had to do with my not outputting the XS tag when aligning with STAR, but that seems not to have resolved.

Perhaps there is a missing BAM tag that MAJIQ needs to read?

Again, would be helpful if the authors had some clarifciation on this point.

[brown.a...@gmail.com]

Dear MAJIQ authors, just wondering if there is an answer to this question?

[Paul Jewell]

Hello All,

We have developed a fix that should work. It is located in the git branch 'strandness-fix'. Could you please check out this branch, reinstall, and try to run the analysis again?

Thank you,

-Paul

[Paul Jewell]

As a follow up, this fix has worked for some others, so I've merged it back into master. You may simply reinstall majiq as before to receive the update.

Thanks.