Can we get absolute abundance of TCR using TCR-Seq?

Question

[CY]

I have a project in which I want to compare the absolute abundance of TCR across patients.

I have no problem deriving relative abundance of each TCR using TCR-Seq. How about absulote abundance?

With DNA barcode, we can calculate the amount of original fragment before PCR. Beside, I can use the same initial blood volumn. However, how can we control the initial DNA input during Library prep (Appearently, we can't use the same amount of DNA input to start because that would be pointless). Is there anyway we can do this without adding spike-in?

[Friederike]

Just out of curiosity: why would it be important to know the absolute abundance? And isn't relative abundance already a touchy topic due to the tiny sampling that we typically do out of the huge universe of possible clonotypes within a person's blood?

[CY]

I gave it another thought and I got a bit confused. The relative abundance of TCR is relative to what? We can't compare the its relative abundance across patients because we got different initial DNA amount across patients, right?

The point is that we are comparing TCR status across patients (undergoing immunotherapy) and aiming to decide the group of patients whose immune systems are boosted in terms of T-cell quantity. Here we think the quantity of TCR is as much important as the diversity of TCR.

[Friederike]

well, hopefully the clonotype that's recognizing the tumor should be boosted within a patient, i.e. it's relative abundance compared to the other clonotypes within the sample from the same patient

[CY]

You are right. We can compare abundance of between clonotypes within an individual. I guess it is not making much sense comparing the total TCR abundance or abundance of specific clonotype between patients (even though in theory we can do it by exacting the same amount of blood and using UMI to keep original amount of DNA fragments)

[Friederike]

Yeah, but it's not really a meaningful measure.

Answer 1

[mizraelson]

I would say the most common setting for that is to normalize your data (to aaccound for different sample size) to the equal number of UMIs / reads / top number of clonotypes etc. and than compare your samples.

You can read on various types of normalization here: https://docs.milaboratories.com/mixcr/reference/mixcr-downsample/

There are also methods that rely on replicas when comparing clonotype fraction between samples. But I don't think you need absolute numbers.