Back

Deviations in Predicted COVID-19 cases in the US during early months of 2021 relate to rise in B.1.526 and its family of variants

Nandakishore, P.; Lui, M.; Prakash, R.; Gourneni, S.; Sukumaran, R.; Berman, J. M.; Iyer, S. A.; Glorioso, C. A.

2021-12-08 epidemiology
10.1101/2021.12.06.21267388 medRxiv
Show abstract

ObjectiveTo investigate the abrogation of COVID-19 case declines from predicted rates in the US in relationship to viral variants and mutations. DesignEpidemiological prediction and time series study of COVID-19 in the US by State. SettingCommunity testing and sequencing of COVID-19 in the US. ParticipantsTime series US COVID-19 case data from the Johns Hopkins University CSSE database. Time series US Variant and Mutation data from the GISAID database. Main outcome measuresPrimary outcomes were statistical modeling of US state deviations from epidemiological predictions, percentage of COVID-19 variants, percentage of COVID-19 mutations, and reported SARS-CoV-2 infections. ResultsDeviations in epidemiological predictions of COVID-19 case declines in the North Eastern US in March 2021 were highly positively related to percentage of B.1.526 (Iota) lineage (p < 10e - 7) and B.1.526.2 (p < 10 - 8) and the T95I mutation (p < 10e - 9). They were related inversely to B.1.427 and B.1.429 (Epsilon) and there was a trend for association with B.1.1.7 (Alpha) lineage. ConclusionDeviations from accurate predictive models are useful for investigating potential immune escape of COVID-19 variants at the population level. The B.1.526 and B.1.526.2 lineages likely have a high potential for immune escape and should be designated as variants of concern. The T95I mutation which is present in the B.1.526, B.1.526.2, and B.1.617.2 (Delta) lineages in the US warrants further investigation as a mutation of concern.

Matching journals

The top 4 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.