todo.org

bcbioRNASeq

Development

Need to handle edge case of kallisto usage with GENCODE.

Need to split based on the pipe and take the first instance for the transcript ID. https://support.bioconductor.org/p/9149475/

Need to improve code coverage / error messages against `ensemblRelease` mismatch.

#190

R Markdown sidebar now is messed up in render output.

Consider reworking AcidMarkdown defaults to address this.

Consider rounding raw counts to integer.

This better matches conventions used in internal bcbio-nextgen R script. Non-integer counts tend to confuse people. Only want to implement this at gene level.

Inform the user in the QC template which aligner was used for QC (e.g. STAR, HISAT2).

Look into tximeta integration.

Consider adding an alert message that fast mode doesn’t support functional analysis.

Recommend that organism, genomeBuild, ensemblRelease are defined.

Improve error message when user passes in incorrect genome.

e.g. Ensembl 105 when using bcbio default hg38 genome. See logs from 2022-04-05 for debugging.

Consider reexporting defunct flatFiles here so we don’t break bcbio 1.2.8.

In bcbioRNASeq class checks, consider allowing NULL tx2gene when level = “transcripts”.

In internal `.tximport` function, need to dynamically set `ignoreTxVersion`?

In internal `.tximport` function, allow NULL tx2gene input when level = “transcripts”.

Need to improve code coverage of kallisto and sailfish import.

Supporting loading of multiple bcbio runs into a single object?

#141

Transcript level analysis with sleuth.

This is supported but we need to improve documentation on how to hand off to sleuth from the bcbioRNASeq object.

DTU with DRIMSeq and DEXSeq.

Need to document how to do handoff from the bcbioRNASeq object for this.

Likelihood ratio test support.

Need to improve documentation and/or add an R Markdown for this. #145

Require valid names in all slots.

Consider renaming “normalized” assay to “sizeFactor” or simply “sf”.

plotMeanSD edgeR RLE is messed up for new LnCap data.

Consider slotting aligned counts into object automatically in updateObject call.

`plotGeneSaturation()`: Consider adding y-intercept 0.

bcbioRNASeq: Improve error message when user doesn’t point to `final/` directory.

Or make this more flexible and detect `final/` automatically.

`plotMappingRate()`: Plot exonic, intronic, intergenic rates.

Show these similarly to qualimap output in MultiQC report.

`plotCountsPerGene()`: Don’t hide the sample legend for `sampleName` interesting groups.

Add RSEM support for tximport? Is this supported in bcbio?

Work on additional multiQC-like plots.

`plotInsertSize()`, `plotDuplicatedReads()`, `plotProperlyPairedReads()`, `plotReadsAligned()`, `plotGCContent()`, `plotXYCounts()`.

`updateObject()`: Add update method for run-length encoding in rowRanges.

F1000 manuscript revisions

Switch paper to use `deg()` instead of `significants()`.

Miscellaneous

Consider running BFG on the repo to remove old binary commits.