In the last few days a number of genomes have been uploaded by South Africa, Australia and Canada that whilst having many of the defining mutations of B.1.1.529 (Omicron) do not have the full set and also have a number of their own unique mutations. This was first described in Issue #359

The mutations are as follows:

where the yellow column gives the mutations unique to the new 'BA.2' lineage, blue column the mutations unique to the original 'BA.1' lineage of Omicron and the green column those shared by both - i.e. inferred to have arisen prior to the split.

A CSV version of this table is available here: Omicron_BA.1_BA.2_mutations.csv

Here we propose expanding the breadth of the B.1.1.529 lineage to include all of these variants. Then 2 sub-lineages created - BA.1 for the original globally-distributed lineage and BA.2 for the new outlier lineage. The names BA.1 and BA.2 follows the Pango convention to avoid more than three numerical fields, an alias is made for the parent lineage.

A couple of observations -

Both sub-lineages (and thus we assume the common ancestor) carry almost all the spike RBD mutations first noted for Omicron and both furin cleaveage adjacent mutations. They both have the NSP6 deletion seen in other VOCs.

The new sub-lineage (putative BA.2) does not carry the spike:69/70del deletion and will thus not be detectable by SGTF (S-gene target failure).

Pangolin currently assign the outlier lineage as B.1.1.529 but Scorpio will give the additional label 'Probable Omicron' because the outlier lineage is missing many of the original defining mutations.