Working with Vierstra motif clusters

[jindalk]

I've been using the vierstra motif PWMs for chromVAR analysis. Is there a way of calculating chromVAR scores on a per motif cluster basis (by somehow merging PWMs of motifs from the same cluster)? Or can I combine chromVAR scores from motifs of the same cluster, after computing Deviations?

One possible solution I could think of is to download the archetype bed file from here https://resources.altius.org/~jvierstra/projects/motif-clustering/releases/v1.0/ , split it by motif cluster and add them as a peak annotation.

Thanks

[rcorces]

I'm not sure I fully understand your question. Can you read this issue post and let me know if this answers it?
#214

[rcorces]

Please let me know if you test this out and it works in your hands. In release_1.0.2 there are some parameter changes, including annoName instead of the old name.

#' @param ArchRProj An `ArchRProject` object.
#' @param motifSet The motif set to be used for annotation. Options include: (i) "JASPAR2016", "JASPAR2018", "JASPAR2020"
#' which gives the 2016, 2018 or 2020 version of JASPAR motifs, (ii) one of "cisbp", "encode", or "homer" which gives the
#' corresponding motif sets from the `chromVAR` package, or (iii) "vierstra" which gives the clustered archetype motifs
#' created by Jeff Vierstra (https://github.com/jvierstra/motif-clustering). 
#' @param annoName The name of the `peakAnnotation` object to be stored in the provided `ArchRProject`
#' @param species The name of the species relevant to the supplied `ArchRProject`. This is used for identifying which motif to be
#' used from CisBP/JASPAR. By default, this function will attempt to guess the species based on the value from `getGenome()`.
#' @param collection If one of the JASPAR motif sets is used via `motifSet`, this parameter allows you to indicate the JASPAR
#' collection to be used. See `getMatrixSet()` from `TFBSTools` for all options to supply for collection. If `motifSet` is
#' "vierstra", then this must either be "archetype" (for the v2 clustered models) or "individual" (for the original v1 individual motif models). 
#' NOTE: vierstra archetype motifs are currently in beta and have not been finalized by Jeff Vierstra.
#' @param motifPWMs A custom set of motif PWMs as a PWMList for adding motif annotations.
#' @param cutOff The p-value cutoff to be used for motif search. The p-value is determined vs a background set of sequences
#' (see `MOODS` for more details on this determination).
#' @param width The width in basepairs to consider for motif matches. See the `motimatchr` package for more information.
#' @param version An integer specifying version 1 or version 2 of chromVARmotifs see github for more info GreenleafLab/chromVARmotifs.
#' @param force A boolean value indicating whether to force the `peakAnnotation` object indicated by `annoName` to be overwritten if
#' it already exists in the given `ArchRProject`.
#' @param logFile The path to a file to be used for logging ArchR output.
#' @param ... Additional parameters to be passed to `TFBSTools::getMatrixSet` for getting a PWM object.

[jindalk]

wow, great work adding it to ArchR natively! Will report back when I try it

[jindalk]

I just tried it and it seems to be working really well! Thanks again for adding this to ArchR. One quick suggestion: Perhaps ArchR could print this URL when vierstra v2 annotation is added? https://resources.altius.org/~jvierstra/projects/motif-clustering-v2.0beta/

When working with archetype results, I would keep going back to this page to make sense of the results. This isn't critical to any functionality, just for added convenience.

[rcorces]

sure. I've added that:

> projHeme5 <- addMotifAnnotations(ArchRProj = projHeme5, motifSet = "Vierstra", collection = "archetype")
ArchR logging to : ArchRLogs/ArchR-addMotifAnnotations-9d612bb3443a-Date-2022-05-06_Time-11-54-16.log
If there is an issue, please report to github with logFile!
2022-05-06 11:54:16 : Gettting Motif Set, Species : , 0.003 mins elapsed.
Using Vierstra v2.1beta motifs. See https://resources.altius.org/~jvierstra/projects/motif-clustering-v2.1beta/ for more details.
Motif file Vierstra_Archetype_Motifs_v2.1.rds does not exist! Downloading..
trying URL 'https://jeffgranja.s3.amazonaws.com/ArchR/Annotations/Vierstra_Archetype_Motifs_v2.1.rds'
Content type 'binary/octet-stream' length 53319 bytes (52 KB)
==================================================

Worth noting that there is already a v2.1beta of the Vierstra motifs due to a bug that was found in their code for the v2.0beta and ArchR has been updated accordingly to use v2.1beta.

[jindalk]

Awesome, thank you!