fix: junctions extractor count overlapping read pairs once#196
fix: junctions extractor count overlapping read pairs once#196TimD1 wants to merge 5 commits intogriffithlab:masterfrom
Conversation
- closes griffithlab#176, CMake was including `src/utils/bedtools/gzstream/version` as a C++ source file, when it was a simple text file containing version info - I moved this version info into `src/utils/bedtools/gzstream/README`
- fixes griffithlab#188, min_intron_length_ was accidentally set from min_anchor_len_
- closes griffithlab#183, flag based filtering - closes griffithlab#189, mapping quality based filtering - option '-F' filters reads containing any of these flags - option '-f' filters reads not containing all these flags - option '-q' filters reads below this mapping quality
- closes griffithlab#186, reads now only 'support' a junction if they have at least a given minimum anchor length, supplied with the '-A' flag (default 0)
- closes griffithlab#187, implemented by adding set of `reads` to `Junction`, and only incrementing `read_count` if read has not been seen yet - this commit should also increase RegTools speed by removing the barcode updating bottleneck caused by repeatedly copying the barcode map - updated regtools to v1.1.0
Dear TimD1, Is the source code for Regtools version 1.1.0 available for download? I am currently using version 1.0.0 and would like to access the updated version so that the Regtools junctions extractor counts overlapping read pairs only once. Thank you for your help. Best regards, |
|
The source code that counts overlapping read pairs only once is available from my fork of RegTools, at: https://github.com/fulcrumgenomics/regtools/tree/187/td/no-double-read-support However, please note that this is not an official RegTools v1.1.0 release (I just suggested bumping the version due to several command-line parameter changes) and there are no guarantees on this fork being supported/maintained (it was only created to make this batch of PRs). |
Hi, Tim, Thank you very much for your help. Does this version address the strandedness issue mentioned in this post (#197)? Xiao |
2 participants
readstoJunction, andonly incrementing
read_countif read has not been seen yetupdating bottleneck caused by repeatedly copying the barcode map