it would be useful for the reader to see the results of these models applied to an emerging variant, and to be able to visually compare estimates at these critical times across models, similar to the approach taken in [Susswein et al](https://www.medrxiv.org/content/10.1101/2023.01.02.23284123v4.full). Figure 1 could be expanded to focus in on say the emergence of a variant that ultimately took off and how the model estimate compared to the ground truth and the data available as of that time, for each reference date, for the MLR model versus a less accurate model.
TODO:
Restrict to two emergences and show the effect of splitting Pango lineages where one emergence from the dominant clade (BA.2 → BA.2 + BA.5, March 1, 2022), one from low fitness clade (BA.2 emergence in July 1, 2022 since sister clade BA.1 was dominant) for one country, one model. Splitting Figure 3 to A B and C where A,B are concrete cases on emergence. Show clade frequencies (as from Nextclade at the time) through time, after solid show. For retrospective data, we need to sum over child clades.
it would be useful for the reader to see the results of these models applied to an emerging variant, and to be able to visually compare estimates at these critical times across models, similar to the approach taken in [Susswein et al](https://www.medrxiv.org/content/10.1101/2023.01.02.23284123v4.full). Figure 1 could be expanded to focus in on say the emergence of a variant that ultimately took off and how the model estimate compared to the ground truth and the data available as of that time, for each reference date, for the MLR model versus a less accurate model.TODO: Restrict to two emergences and show the effect of splitting Pango lineages where one emergence from the dominant clade (BA.2 → BA.2 + BA.5, March 1, 2022), one from low fitness clade (BA.2 emergence in July 1, 2022 since sister clade BA.1 was dominant) for one country, one model. Splitting Figure 3 to A B and C where A,B are concrete cases on emergence. Show clade frequencies (as from Nextclade at the time) through time, after solid show. For retrospective data, we need to sum over child clades.