T Cell Receptor Sequencing total reads/clone count vs unique clones

Question

[bioinformatics.cancer]

Hi,

I have a pre and post immunotherapy (oncolytic virus) T Cell Receptor (DNA) sequencing data set that I am analyzing to draw some conclusions on the effect of treatment on the T Cell receptor repertoire, such as finding expansion of specific clones after treatment and change in clonality. TCR sequencing was done with blood (PBMC) samples and not tumor samples. The data that I have includes DNA and AA sequences of each clone in a sample repertoire and the total number of clones, along with other details such as the V and J gene usage. The number of unique clones and the total number of clones vary a great deal between samples (pre and post treatment) and in general tend to be much higher after treatment. I am not sure why the total number of clones varies so much since the same amount of blood sample was drawn. It could be due to variation in the number of mature T Cells between patients and as a result of treatment. I have done several types of analyses but was wondering if there is some information I can extract from the total number of reads/clones and total number of unique clones, such as the ratio of the two. Thank you for your help.

• Pankaj

Answer 1

[mikhail.shugay]

You can check out this paper of mine for some ideas on normalization techniques for estimating T-cell repertoire diversity, and try out VDJtools

A benchmark of diversity estimators is given in the supplementary. Feel free to ask any questions regarding this paper here.

Answer 2

[Charles Plessy]

The ratios of unique versus total number of clones of one library will decrease when sequencing depth increases. Be careful with this, because you can only use this value to compare libraries that had a similar total number of reads. The solution to that problem is to chose a diversity index that is independent from the sequencing depth. I am very satisfied of the richness index implemented in the rarefy function of the R package "vegan".