search

cov-lineages / lineages-website

16 stars 13 forks source link

feat: add covSPECTRUM to dashboards #7

Closed corneliusroemer closed 3 years ago

corneliusroemer commented 3 years ago

covSpectrum is superior to both outbreak.info and covariants.org in terms of querying options.

For example, a pango lineage can be queried with a wildcard, including all descendant lineages: B.1.617.2* also includes all AY*.

It is also updated with new data much faster (within hours of new GISAID data) than outbreak.info and covariants (which take 1-3 days).

For further information contact @chaoran-chen

Declaration of Interests:

emilyscher commented 3 years ago

Thanks for this. By the way, it looks like you're conflating lineages with variants. It might be worth taking a look at that.

corneliusroemer commented 3 years ago

Thanks for merging and the tip about lineage vs. variant.

Can you point me to a definition? CDC says "variant, lineage, strain" are used interchangeably. Couldn't find anything authoritative that pointed out distinctions. image

https://web.archive.org/web/20210122060907/https://www.cdc.gov/coronavirus/2019-ncov/more/scientific-brief-emerging-variant.html

rambaut commented 3 years ago

Essentially if you are referring to the Pango designated labels, 'B.1.1.7' etc then these should be referred to as lineages or Pango lineages. Variant is a much vaguer term and should probably be avoided (my preference is just to use it when talking about genetic variation specifically). Variants of Concern/Interest are obviously a WHO designated thing and are probably best referred to as VOC/VOI (i.e., Alpha, Delta etc). Noting that because the 'concern' is for the mutation constellation, any descendant sublineage which also has this constellation should still be referred to as that VOC/VOI (i.e., any AY.x lineage is a Delta VOC).

rambaut commented 3 years ago

Many virologists consider SARS-CoV-2 to be the strain (and I would agree).

aineniamh commented 3 years ago

https://www.cogconsortium.uk/what-do-virologists-mean-by-mutation-variant-and-strain/

Hi @corneliusroemer, COG-UK has a good explainer on these terms and I know theres some conflict info out there! If you're using a name that fits in the Pango nomenclature scheme it's a lineage, rather than a variant!

corneliusroemer commented 3 years ago

Thanks @rambaut and @aineniamh for explaining.

Regarding the explainer: It defines the term variant as: "A variant is the whole sequence of the virus (the genome), which may contain one or more mutations." which I understand as a unique sequence. An identical sequence from a different person would be the same variant. If the sequence differs by any amount, even just a single nucleotide mutation it'd be a different variant.

If you use that definition of a variant, then lineages are a set of one or more variants (that share certain mutations).

Confusingly, the COG explainer is internally inconsistent. It first defines what a variant is, but then uses the term variant in a way that is inconsistent with this definition when talking about VOCs: "COG-UK, in partnership with the public health authorities, carefully track all interesting variants, and if there is clear evidence that the variant is causing problems then the variant is called a Variant of Concern."