First day of qImmunology workshop: diversity and error

It's the first day of the quantitative Immunology workshop in Les Houches (#qImmLH), and there's been a definite theme: TCR repertoire sequencing. In fact it seems to be the main theme of the conference, with around half the people here seemingly working on them in one sense or other – my accommodation alone seems to be populated exclusively with us*! Seeing as I have a bit of time today, have a quick post about it.

There's been a couple of recurrent points which have been coming up in the talks, and are being particularly dwelt on by the discussions from the group (being based in a Physics retreats apparently demands that we must do as the physicists do, and ask questions throughout a talk). Well, there have been many but these are the two I picked up on most, but that might just be because it's what my poster is about so I'm biased.

The first is that of error. So far, all of our pipelines involve some amount of PCR amplification, which adds a great deal of errors on top of whatever the sequencing technique itself will introduce. My own supervisor Benny Chain probably dwelt on this the longest, going over some evidence to suggest that within a given PCR there's some variability in the efficiency of amplification, so for lower frequency clones there's less relationship between the number of reads coming out and the number of original molecules of DNA that went in. However a number of the talks touched on this, and I'm sure some of the later ones will as well.

The second theme is that of diversity, and how to measure it. Based on the fact all of the speakers used a different metric (and the number of questions it raised from the audience) there's clearly scope for discussion. In brief, Encarnita Mariotti-Ferrandiz used a species richness index to describe the number of different unique clonotypes in mice Treg and Teff cells, Thierry Mora used Shannon Entropy to compare diversities of zebrafish Ig CDR3s, and Eric Shifrut looked at Gini indexes of aging mice repertoires.

This of course all ties in to the error of the system, as any additional error will be likely be artificially inflating the diversity while at the same time distorting the frequency distribution.

The last full talk of the day ended with Aleksandra Walczak talking through the generation of diversity in TCR repertoires, mostly just going through the figures from the excellent Murugan et al paper.

So far this is shaping up to be an ideal workshop for me, what with so many people working on and talking about the exact problems I find myself faced with on a daily basis. That it's all taking place among some achingly beautiful scenery is just icing on the cake.

* I know it's an awful thing to think about, but I can't help feeling that if an avalanche hit the resort we'd be setting the relatively young field of adaptive repertoire sequencing back a decent way!