Data overlap after adding rnaseq data from literature

[dktanwar]

Overlap and perform analysis as in #28

Changes

• RNA-seq data is from literature.
• rna_cutoff: Instead of using (lfc > 1 & qval < 0.05), now we only use (lfc > 1).

Results directory

Folder ending with the issue number (-36).

GO analysis using GREAT

Sub-folder 05. Further, in the output folder.

TF motif enrichment analysis using HOMER

Sub-folder 06. Further, in the output folder.

For further details on the results organization: #28

Enriched Heatmaps

The enriched heatmaps are created:

1. Without including ChIP-seq data
2. Without including ChIP-seq data (log₂)
3. Including ChIP-seq data
4. Including ChIP-seq data (log₂)

Heatmaps

• [x] Distal regions
• [x] Distal regions with more accessibility in Adults
• [x] Distal regions with less accessibility in Adults
• [x] Proximal regions
• [x] Proximally active regions
• [x] Proximally inactive regions

Heatmaps directory: 20201001-04_EH_DO

logFC for RNA-Seq (colors)

sqrtCols <- function(x, cols = NULL) {
  if (is.null(cols)) {
    cols <- rev(brewer.pal(n = 11, "RdBu"))
  }
  if (is.function(cols)) {
    cols <- cols(11)
  }
  if (length(cols) != 11) {
    stop("`cols` should contain 11 colors.")
  }
  bb <- c(-seq(from = sqrt(abs(min(x, na.rm = TRUE))), to = 0, length.out = 6)^2, 
          seq(from = 0, to = sqrt(max(x, na.rm = TRUE)), length.out = 6)[-1]^2)

  colorRamp2(bb, cols)
}

Received this code from @plger!

[dktanwar]

@Irinalazar

I checked my code and analysis is running. I will further check it after the results are available.

[dktanwar]

As per discussion on 09.10.2020, we agreed to also perform overlap analysis within a window of 1kbp around mid of the peak regions.

For better visualization, the number of windows on the heatmap is the same (50).

Also, @plger suggested creating Enriched Heatmaps without methylation data. As we are not talking much about the methylation data, figures with methylation data can go into supplementary (that we made use of methylation data from literature).

[dktanwar]

Further, for the methylation data, I also asked the author of Enriched Heatmap about his opinion. The code we are using seems to be fine. https://github.com/jokergoo/ComplexHeatmap/issues/600