Cumulative Prevalence also consider sequences without pango lineage assigment

[fabou-uobaf]

Dear all

I just realized that the Cumulative Prevalence for different lineages in your superb web resource seems to be calculate by #GISAID_sample_with_lineage_X/#GISAID_sample. Generally this seems to work fine, but in the case of Austria, where there were recently submitted 20K samples of partial genome sequencing to GISAID (~80% of all samples). Since those do not have a pango lineage assigned, in the case for Austria a calculation like #GISAID_sample_with_lineage_X/#GISAID_sample_with_any_lineage would be more appropriate.

bw, Fabou

[gkarthik]

Hello Fabou, thank you for pointing this out. We do have a length filter step that excludes sequences under 20,000 bps but in this case, it seems to have failed. I will look into this. Excluding low length sequences should fix the prevalence estimates too.

[AlaaALatif]

hello, a fix for this has been added as part of this PR: andersen-lab/bjorn#15

It should be reflected on outbreak.info once we merge and update our production pipeline.