Create a geneactivity heatmap of custom genes and custom clusters?

[cschmidl]

I would like to plot a custom heatmap (like the marker gene heatmap) but with a specific list of genes and clusters. Can anyone help? I tried to subset the SummarizedExperiment derived with getMatrixFromProject or built with getMarkerFeatures but I could not get it done. Thanks!

[rcorces]

> markerTest <- getMarkerFeatures(
+   ArchRProj = projHeme5, 
+   useMatrix = "GeneScoreMatrix",
+   groupBy = "Clusters2",
+   testMethod = "wilcoxon",
+   bias = c("TSSEnrichment", "log10(nFrags)"),
+   useGroups = "Erythroid",
+   bgdGroups = "Progenitor"
+ )
ArchR logging to : ArchRLogs/ArchR-getMarkerFeatures-365f6783727c-Date-2021-07-12_Time-08-04-38.log
If there is an issue, please report to github with logFile!
2021-07-12 08:04:38 : Matching Known Biases, 0.001 mins elapsed.
2021-07-12 08:04:39 : Computing Pairwise Tests (1 of 1), 0.017 mins elapsed.
###########
2021-07-12 08:05:05 : Completed Pairwise Tests, 0.453 mins elapsed.
###########
ArchR logging successful to : ArchRLogs/ArchR-getMarkerFeatures-365f6783727c-Date-2021-07-12_Time-08-04-38.log
> markerTest
class: SummarizedExperiment 
dim: 23127 1 
metadata(2): MatchInfo Params
assays(7): Log2FC Mean ... AUC MeanBGD
rownames(23127): 1 2 ... 23126 23127
rowData names(6): seqnames start ... name idx
colnames(1): Erythroid
colData names(0):

> rowData(markerTest)
DataFrame with 23127 rows and 6 columns
      seqnames     start       end  strand         name     idx
         <Rle>   <array>   <array> <array>      <array> <array>
1         chr1     11874     14409       1      DDX11L1       1
2         chr1     29961     14362       2       WASH7P       2
3         chr1     69091     70008       1        OR4F5       3
4         chr1    140566    134773       2    LOC729737       4
5         chr1    714068    700245       2 LOC100288069       5
...        ...       ...       ...     ...          ...     ...
23123     chrX 154299695 154351349       1        BRCC3     906
23124     chrX 154444701 154468098       1         VBP1     907
23125     chrX 154493852 154487526       2       RAB39B     908
23126     chrX 154563986 154505500       2        CLIC2     909
23127     chrX 154842622 154718673       2        TMLHE     910

> length(unique(rowData(markerTest)$name))
[1] 23127

> subsetSE <- markerTest[which(rowData(markerTest)$name %in% c("GATA1","GATA2","GATA3"))]
> subsetSE
class: SummarizedExperiment 
dim: 3 1 
metadata(2): MatchInfo Params
assays(7): Log2FC Mean ... AUC MeanBGD
rownames(3): 2369 15802 22455
rowData names(6): seqnames start ... name idx
colnames(1): Erythroid
colData names(0):
> rowData(subsetSE)
DataFrame with 3 rows and 6 columns
      seqnames     start       end  strand    name     idx
         <Rle>   <array>   <array> <array> <array> <array>
2369     chr10   8096667   8117164       1   GATA3      56
15802     chr3 128212030 128198265       2   GATA2     812
22455     chrX  48644982  48652717       1   GATA1     238

[cschmidl]

Thank you Ryan, I tried this but when then plotting it with plotMarkerHeatmap only a fraction of my selected genes are plotted (even when I do not apply FDR and log2fc cut-offs in the plotMarkerHeatmap function by setting them to high/low values that should not exclude any gene).

[rcorces]

I dont have this problem. Perhaps you can create a reproducible example using the tutorial dataset.