Thread by @LauringLab

Happy to share out latest preprint on SARS-CoV-2 within host diversity. Brief thread.
biorxiv.org/content/10.110…

We've been interested in within-host diversity for some time. It's important for understanding how mutations accumulate over time (relevant for how variants arise). Important for understanding how likely it is for variants to transmit and potentially for inferring trx chains 2/x

We looked at inpatients and outpatients. First we measured viral load. Why? Cause we're interested. But it's also a big factor in looking at diversity within hosts (spoiler alert). Another important (unique?) thing in our study is we analyzed data by day from symptom onset 3/x

We then did a mixing study with variants at set frequencies (e.g. 10%, 5%, 2%) and varying input viral loads (10e4, 10e3) to benchmark our variant identification. As we've shown before, at lower viral loads, lots of false positive variants. This can be major issue in studies 4/x

So figured out a way to screen out false positives. We found that diversity is quite low (lower than some other reports). Most had only 1-3 minority variants at a 2% threshold. 5/x

A few had lots of minor variants. We're not sure why (have to read the preprint to see what we thought about). In general, there weren't that many variants over the time frame when most people transmit (so timed analysis key!). And most are <10% frequency. 6/x

Means it is likely uncommon for newly arising within-host variants to transmit (unless the bottleneck is really big). 7/x

Are variants shared among hosts? Sometimes. We found convergence in rare variant fraction across individuals - who were *unlikely* to be linked by transmission. So perhaps virus exploring similar solutions. These variants did not predict future consensus variants on GISAID 8/x

Are minority variants useful for transmission inference? Perhaps, but use with caution. We found shared variants in pairs with identical and non-identical consensus sequences. Some of whom did not have epi linkage. 9/x

Take home points: (1) QC is critical in within-host diversity studies. (2) Time-based analyses of diversification are informative. (3) No difference in diversity between inpatients and outpatients. (4) Challenges for studies of transmission.

Last one (sorry stopped numbering). Work of @alvalesano. Lots of work from team. @itswillfitz @Kalee_Rumfelt @jgpetrie @MartinEpi Derek Dimcheff, Chris Blair.