multiome analysis

[tanasa]

Hi Ryan,

For Multiome analysis, shall we use the current tutorial that presents the data integration in the chapter :

https://www.archrproject.com/bookdown/defining-cluster-identity-with-scrna-seq.html

or shall we follow a relatively old script that is described in :

https://greenleaflab.github.io/ArchR_2020/Ex-Analyze-Multiome.html.

Thank you !

[rcorces]

Just to clarify - Multiome is very different from integration.
Multiome is two datasets acquired simultaneously from the same cell. By definition, these dont require "integration" because you know which cell they came from.
So this is the multiome tutorial - https://greenleaflab.github.io/ArchR_2020/Ex-Analyze-Multiome.html
The other link you supplied is for when you have ATAC and RNA data acquired separately and you wish to combine/integrate them

[tanasa]

Thank you Ryan. It is always very impressive, as your answers are so prompt.

If I am add a novice question please : in the absence of scRNAseq data, could we perform all the analysis that are described in the ArchR manual only based on scATAC-seq data ?

[rcorces]

Any analysis based on RNA data is, of course, not possible without RNA data. Many manual sections utilize RNA data.

[tanasa]

Thank you Ryan ! We may have a time course scATAC-seq datasets. Which analytical method shall we use ?

[millerkaplanlab]

Is there another tutorial for more detailed downstream analysis for mutliome datasets. I've processed my data through the pipeline; however, I am looking to identify various gene markers in my dataset....

[rcorces]

no real reason. they are quite correlated though

[brian-1111]

no real reason. they are quite correlated though
@rcorces
So in other words using bias = c("TSSEnrichment", "log10(nFrags)", "log10(Gex_nUMI)", "log10(Gex_nGenes)") in getMarkerFeatures is not necessary to produce a more accurate result is that correct?

[brian-1111]

my serna is built like this

seRNA2<-SummarizedExperiment(assays=list(counts=seurat.object)) which three different origin.idents

What is the correct format i do not found it? thanks for your answer

You should follow the tutorial @rcorces provided here https://github.com/GreenleafLab/ArchR_Website/blob/ui_updates/bookdown/21_MultiomicDataAnalysis.Rmd.

import10xFeatureMatrix will input your RNA data as RangedSummarizedExperiment that you can used to merge with your ArchR project

seRNA <- import10xFeatureMatrix(
  input = rnaFiles,
  names = names(rnaFiles),
  strictMatch = TRUE
)

[rcorces]

So in other words using bias = c("TSSEnrichment", "log10(nFrags)", "log10(Gex_nUMI)", "log10(Gex_nGenes)") in getMarkerFeatures is not necessary to produce a more accurate result is that correct?

I cant be sure as I havent tested this

[brian-1111]

So in other words using bias = c("TSSEnrichment", "log10(nFrags)", "log10(Gex_nUMI)", "log10(Gex_nGenes)") in getMarkerFeatures is not necessary to produce a more accurate result is that correct?

I cant be sure as I havent tested this

@rcorces thnaks for your reply. But is there a way I can verify it with my own results?
I'm actually quite new to mitiomics analysis. But based on my knowledge I think using nGenes instead of nUMI for bias input could potentially lead to a more accurate result since nGenes could represent the actual uniquely identified genes. I am curious that why do you consider the UMI to be the only input for adjusting the via of the dataset? Is it because the tutorial is only for the demonstration only?

[ArutStory]

Thank you, you really helped!