MonashBioinformaticsPlatform / RNAsik-pipe

RNAsik - more than just a pipeline
https://monashbioinformaticsplatform.github.io/RNAsik-pipe/
Apache License 2.0
13 stars 5 forks source link

Backward compatible #10

Closed stu2 closed 6 years ago

stu2 commented 6 years ago

A few improvements to how -refFiles directory is used, mainly for use with -align starWithAnn. This now differentiates between the situation where > 1 file in a file-type is give (now errors out, this was previously not tested for) and the situation where not all the 6 filetypes are given. As of the last commit of this branch, this situation just gives a warning, as many pipelines don't need all 6 filetypes searched for, and attempts to generate .fai, .chromSizes and .dict files from the .fa file.

If index is made then the folder name will have .starWithAnnIdx or .starIdx for -align starWithAnn and align star, respectively, allowing the folder to be properly identified next time if the align mode is appropriate to the index.

serine commented 6 years ago

that's right, you don't always need all 6, or however many reference files to go ahead, but -refFiles feature is specifically to reuse previously generated reference files by RNAsik, that is to reuse refFiles/ directory with its content. Any complete RNAsik run will produce that many reference files. At the end of the day the most time consuming step in reference files generation is generating aligner specific index. And if you already have your index that you'd like to reuse, simply use -genomeIdx option to pass that through.

I took bits from this pull request and will push them to master soon, but I'll close this PR. However I really appreciate time spend on this.