Integrate transcriptomics and CNV

Question

[QX]

Dear Community,

I have two omics datasets (WGS and transcriptomics) on tumor/normal cells. I have performed CNV calling on genomics data and DEG on transcriptomics data. Is there any methods/ approach to compare among these two dataset?

Best,

Answer 1

[jared.andrews07]

Depends on your ultimate question(s). A good ol' scatter plot with copy number on one axis and expression on the other for each gene can be useful. Typically, they correlate to some degree.

[QX]

HI jared.andrews07 ,

Thank for your response. That could be a good look at the data.

However, the CNVs are set based on bin size (10kb or 100kb) on reference (whole) genomes, while genes are varying a lot in lengths. For e.g. if a CNVs is defined as 10kb, it would possibly contain several <2kb genes, but only a part of 30kb genes. In addition, a CNVs can possible cover part of the genes, for e.g. start position at 500 of the genes instead of 1 or in front of the gene. As they are not perfect match the position on genomes, I found it difficult to correlate/plot the CNVs vs genes expressions on scatter plot.

How do you think?

[jared.andrews07]

A good question, without a perfect answer, but one option is to use something like cnvkit's genemetrics command to call absolute copy number for each gene, which attempts to use some sensible defaults to deal with those that span multiple segments:

Where more than one segment overlaps the gene, i.e. if the gene contains a breakpoint, each segment’s value will be reported as a separate row for the same gene. If a large-scale CNA covers multiple genes, each of those genes will be listed individually.

In such cases, one could remove it from the scatter plot (or distinguish it with a different color or shape).

[QX]

sorry for late response @jared.andrews07,

you did suggest great way to look at this kind of problem and I am working on this with your approach.

Thank a lot,

Best