Umap

[Rachel Rosenthal]

Hello,

I was wondering if you had any experience trying to apply Umap to a transcriptome, or if this is even necessary. Would the same default Umap bed files be suitable to determine mappability of RNAseq reads?

Thanks,

Rachel

[Mehran Karimzadeh]

Hi Rachel,

Thank you for your interest in Umap.

I haven't applied Umap to transcriptome data.

From my experience, different software take different approaches and there is a strong literature comparing the available tools.

This includes the algorithm for alignment itself as well as how to include multi-mapping reads in gene expression quantification.

We discuss these briefly in the manuscript which you can take a look.

Best,

Mehran

--
You received this message because you are subscribed to the Google Groups "Ubismap" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ubismap+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ubismap/7c7949ef-8fec-4547-a0a1-53091ef3c00en%40googlegroups.com.

[Rachel Rosenthal]

Hi Mehran,

I had a quick question about the "Convert numeric vectors to BED and Wiggle" step. 

I have run the uint8_to_bed.py script, specifying:

<MergedKmerDir> 

<OutDir> 

<LabelForOutputFiles>

But get an error:

KeyError: 'SGE_TASK_ID'

Does this also have to be submitted as an array job, and if so, would it be similar in format to how "combine_umaps.py" was submitted in the previous step?

e.g.

JOBID=$(qsub -q byslot.q -N ToBED -terse -t 1-188813 -cwd -v PATH -b y -o out.LOG -e out.ERR uint8_to_bed.py kmerDir outDir Label -bed -chrsize_path chrsize.tsv -var_id SGE_TASK_ID)

Thanks,

Rachel