Predicting A/B from GC content?

[TerezaClarence]

Dear GENOVA developers,

I was wondering if it is possible to allow A/B compartment prediction and saddle plot analysis by using the GC content (which could be in a *bw format)? If that is true, can I kindly ask for suggestion how to do that or what other format of GC content should be loaded (I assume *bed format should be ok)?

Thank you very much in advance for your help.

Best regards, Tereza

[TerezaClarence]

Dear GENOVA developers,

I was wondering whether you plan to add a possibility of GC-content for creating saddle_plots (re my previous post)? When comparing for instance H27K3ac and GC content for generating saddle_plots in cooltools, we do observe a difference and we may prefer to use GC-content.

Thank you very much in advance for answer.

Best wishes, Tereza

[teunbrand]

Hello there,

Yes it is possible to use GC content for signing the compartment scores, it is alluded to in the documentation of ?sign_compartmentscore. The format of the GC bias should be in that of a bedgraph. An example of how one could use the GC content is below.

library(GENOVA)
library(BSgenome.Hsapiens.UCSC.hg19)
library(Biostrings)

# Get test data
exp <- get_test_data("150k")

# Get relevant chromosome lengths
chroms <- seqinfo(BSgenome.Hsapiens.UCSC.hg19)[exp$CHRS]

# Cut chromosomes into bins
bins <- tileGenome(seqlengths(chroms), tilewidth = 10e3, cut.last.tile.in.chrom = TRUE)

# Retrieve sequence and calculate GC bias
seq <- getSeq(BSgenome.Hsapiens.UCSC.hg19, bins)
seq <- alphabetFrequency(seq, as.prob = TRUE)
gc_bias <- seq[, "G"] + seq[, "C"] 

# Reformat as bedgraph
bedgraph <- data.frame(
  chrom = decode(seqnames(bins)),
  start = start(bins),
  end   = end(bins),
  value = gc_bias
)

# Compute compartment score
cs <- compartment_score(exp, bedgraph = bedgraph)
visualise(cs, chr = 'chr21')
#> Warning: Removed 62 row(s) containing missing values (geom_path).

^{Created on 2021-10-19 by the reprex package (v2.0.0)}

[hiroyukikato911]

Hi,

Thanks for this wonderful tool!

I'm also looking for GC content sorting for AB compartments. I realize that the aforementioned command cannot apply to .hic files that have chromosome name "1", "2" and so on. (without chr) Is there a way to overcome this?

Best regards, Hiroyuki

[teunbrand]

Hi there,

If your HiC data chromosomes are denoted with "1", "2" etc, doesn't it simply solve the problem by gsub()ing out the "chr" part of the chromosome names?

Best, Teun