1/ Using data of 21 #COVID19 variants from 42 countries in a multilevel multinomial model I compared the growth rate of these variants.
The estimated growth advantage of Delta compared to Alpha is 62%/week (95% credible interval: 54-71%).
The estimated growth advantage of Delta compared to Alpha is 62%/week (95% credible interval: 54-71%).
2/ Extrapolation of these numbers for 6 most prevalent variants in 42 different countries.
Left: linear scale
Right: logit scale
Left: linear scale
Right: logit scale
3/ On 19 July, at least 13 countries are expected (with a probability of 97.5%) to have the Delta variant in >90% of the samples.
4/ Something strange seems to happen in Greece where the group of 'other' variants seem to have a similar growth advantage as the Delta variant compared to the Alpha variant. After 14 June no data about variants is available for Greece.
5/ Source data: github.com/hodcroftlab/co…
In comparison with previous threads, I excluded Canada and Indonesia because the sum of variants was larger than the total number of sequences @firefoxx66
See also @TWenseleers who's fitting similar models
In comparison with previous threads, I excluded Canada and Indonesia because the sum of variants was larger than the total number of sequences @firefoxx66
See also @TWenseleers who's fitting similar models
https://twitter.com/TWenseleers/status/1412280640322539520
• • •
Missing some Tweet in this thread? You can try to
force a refresh
