search

qqwang-berkeley / JUM

A tool for annotation-free differential analysis of tissue-specific pre-mRNA alternative splicing patterns
MIT License
27 stars 11 forks source link

AS junction co-ordinates are sometimes wrong. #29

Open angel-bee2018 opened 4 years ago

angel-bee2018 commented 4 years ago

Hi,

I've been using the time-series pipeline of JUM and noticed that the junction co-ordinates are consistently lower compared to when viewed in IGV as well as the reference annotation. This problem seems to begin all the way from the JUM_temp_A step, with the UNION_junc_coor co-ordinates already different to those in the SJ.out.tab file from STAR.

Upon further exploration, I noticed that the co-ordinates are usually 1 less than expected. Under certain circumstances (mostly in cassette exon calls), some co-ordinates may be correct and some may be short between sub-junctions with in the same AS junction.

I hope that this issue can be looked into. Best wishes

angel-bee2018 commented 4 years ago

I am more than happy to provide additional files to help pinpoint this discrepancy.

angel-bee2018 commented 4 years ago

AS_event_ID as described by JUM: image Illustration in IGV. green represents co-ordinates matching with that from STAR. red represents discrepancy. edit: i forgot to mention the set of transcript annotations below the reference is a reconstructed transcriptome from my data. it is not the reference annotation. image Snippet of SJ.out.tab It seems that the .bam junctions view in IGV has minused all co-ordinates by 1 across the board. I initially thought it had something to do with co-ordinates taken as starting from 1 compared to 0. However, the additional short-fall in red highlighted co-ordinate in IGV cannot be explained. image

rLannes commented 1 year ago

This may be due to difference in sam (1 indexed)/bed(0 indexed) format indexing. For more information, I found this discussion: https://www.biostars.org/p/84686/.

Do you know if it has been fixed?