What is the real meaning of homozygous in the tumor but heterozygous in normal samples?

Question

[mangfu100]

Hi.

I have read paper VarScan2 : somatic mutation and CNA discovery today.

In the paper, I didn't understand the statement below.

VarScan performs a direct comparison between tumor and normal genotypes and supporting read counts to determine the somatic status. Variants present in both samples are classified as somatic, variants heterozygous in the normal but homozygous in the tumor are classified as LOH, and variants shared between samples are classified as germline.

The bold texts is that I can't understand the concept.

Firstly, I don't know the difference between germline and somatic considering the context above. I have already knew that germline is that varaints that also exist in tumor and normal but somatic is that variants only exist in tumor. but above sentence is so confused. I think the author wrote the sentence not clear. Is that right?

Secondly, I don't understand variants classified as LOH. Is it really important to classify variants considering homozygous and heterozygous? I don't know the reason why VarScan2 classify according to homozygous variants in tumor sample. What is the real meaning of homozygous in the tumor but heterozygous in normal samples? Is that important to consider? I don't know..

My background is computer science not biology, so homozygous and heterozygous concept are always very hard to understand.

I hope you understand.

Thank you!

Answer 1

[Devon Ryan]

Since you've asked questions like this a couple times, you might find a local biologist to ask. These end up not really being bioinformatics questions, though I don't know of a good place to ask online.

The sentence is clear to me, but being an overly-educated native speaker that doesn't mean much. I'll reword what's written and try to provide a bit of context in doing so:

Determining variants likely involved in cancer biology is a difficult process. One reason for this is that we don't have a perfect reference genome for each patient that we encounter. Consequently, although one can find a plethora of variants in their tumour samples, most of these aren't biologically relevant. Our task, then, is to separate the biologically relevant from the biologically irrelevant variants. One way to do that is to determine if a variant is germline or somatic. If a variant is germline then it was inherited (or otherwise transmitted) from one of a patient's parents. It will then be present in both the tumour sample and normal tissue. Because of this, it's unlikely to be of prime interest. A somatic variant, on the other hand, arises due to a mutation at some point in development/adulthood. If these are related to the cancer, they'll be present in the cancer sample but not the surrounding tissue.

For LOH, I'll just link to the wikipedia article. Suffice it to say, if you have a sudden loss of 50% of something (assuming an unbalanced event) then you'll likely have issues because of it. We see similar things happen in daily life when roads are being worked on. If it's a small road there's not much of a problem, but a 50% or so loss in capacity of a major road can cause huge problems. The genome isn't so different in this regard.

[mangfu100]

Thank you.

I am very appreciate your answer :)