Mercurial > repos > fubar > bigwig_outlier_bed
diff README.md @ 0:c71db540eb38 draft
planemo upload for repository https://github.com/jackh726/bigtools commit ce6b9f638ebcebcad5a5b10219f252962f30e5cc-dirty
| author | fubar |
|---|---|
| date | Mon, 01 Jul 2024 02:48:46 +0000 |
| parents | |
| children | eb17eb8a3658 |
line wrap: on
line diff
--- /dev/null Thu Jan 01 00:00:00 1970 +0000 +++ b/README.md Mon Jul 01 02:48:46 2024 +0000 @@ -0,0 +1,31 @@ +## bigwig peak outlier to bed + +### July 30 2024 for the VGP + +This code will soon become a Galaxy tool, for building some of the [NIH MARBL T2T assembly polishing](https://github.com/marbl/training) tools as Galaxy workflows. + +The next JBrowse2 tool release will include a plugin for optional colours to distinguish bed features, shown being tested in the screenshots below. + +### Find and mark BigWig peaks to a bed file for display + +In the spirit of DeepTools, but finding contiguous regions where the bigwig value is either above or below a given centile. +0.99 and 0.01 for example. These quantile cut point values are found and applied over each chromosome using some [cunning numpy code](http://gregoryzynda.com/python/numpy/contiguous/interval/2019/11/29/contiguous-regions.html) + + + + + +Big differences between chromosomes 14,15,21,22 and Y in this "all contigs" view - explanations welcomed: + + + + +[pybedtools](https://github.com/jackh726/bigtools) is used for the bigwig interface. Optionally allow +multiple bigwigs to be processed into a single bed - the bed features have the bigwig name in the label for viewing. + +### Note on quantiles per chromosome rather than quantiles for the whole bigwig + +It is just not feasible to hold all contigs in the entire decoded bigwig in RAM to estimate quantiles. It may be +better to sample across all chromosomes so as not to lose any systematic differences between them - the current method will hide those +differences unfortunately. Sampling might be possible. Looking at the actual quantile values across a couple of test bigwigs suggests that +there is not much variation between chromosomes but there's now a tabular report to check them for each input bigwig.