Splice types #215
Splice types #215
Conversation
| #' - theta: '`sm cfg.getProcessedDataDir()+ | ||
| #' "/aberrant_splicing/datasets/savedObjects/raw-{dataset}/theta.h5"`' | ||
| #' - txdb: '`sm cfg.getProcessedDataDir() + "/aberrant_expression/{annotation}/txdb.db"`' |
There was a problem hiding this comment.
Snakemake cannot find the txdb, because the annotation wildcard is not specified in the output.
I wouldn't annotate the fds here anyway, because it is not thought to be annotation-depedent. Rather, it would be more useful to only annotate in the steps after FRASER fitting
| #' - spliceTypeSetup: '`sm cfg.AS.getWorkdir() + "/spliceTypeConfig.R"`' | ||
| #' - addAnnotation: '`sm cfg.AS.getWorkdir() + "/fds_annotation.R"`' | ||
| #' - addSpliceType: '`sm cfg.AS.getWorkdir() + "/spliceType_frameshift_annotation.R"`' | ||
| #' - subtypes: '`sm cfg.AS.getWorkdir() + "/subtypes_exonSkipping_inconclusive.R"`' | ||
| #' - annotate_blacklist: '`sm str(projectDir / ".drop" / "helpers" / "annotate_blacklist.R")`' |
There was a problem hiding this comment.
I would strongly advice to have this in a separate package or within FRASER. But not as just scripts lying around within drop. This functionality could be even good for FRASER by itself. So please think twice if you want to put it just here as is.
| #' - ids: '`sm lambda w: sa.getIDsByGroup(w.dataset, assay="RNA")`' | ||
| #' - fileMappingFile: '`sm cfg.getRoot() + "/file_mapping.csv"`' | ||
| #' input: | ||
| #' - setup: '`sm cfg.AS.getWorkdir() + "/config.R"`' |
There was a problem hiding this comment.
Are we sure we want the config in requirements and not as params? There should be nothing in there that could change the outcome of the results.
This goes of course also for the other scripts.
|
|
||
| # Add the junction annotations to the fds | ||
| #message("03: load db for annotation") | ||
| #txdb <- loadDb(snakemake@input$txdb) |
There was a problem hiding this comment.
Please clean up if you do not need the lines.
|
|
||
| ### Do the annotation just for the most used intron (highest median expression) | ||
| print("start calculating annotations") | ||
| annotations <- sapply(c(1:length(fds_junctions)), function(i){ |
There was a problem hiding this comment.
not sure how fast this is, but rather think of vectorized versions. granges is optimized to find overlaps between 1000s of ranges but not 1 by 1 over 1000s ranges.
| @@ -0,0 +1,67 @@ | |||
| ### 20210604 klutz | |||
|
|
|||
| ### basic annotations (start, end, none, both) for full fds | |||
There was a problem hiding this comment.
Add a bit more details what the function is doing. Especially going through it you somehow only looking at the max expressed junctions? Why? Please describe this as a doc. For more details see roxygen and some bioconductor packages.
|
Hi, is this PR still being worked on? |
Annotate Splicing outputs with different splice types.