search

andersen-lab / Freyja

Depth-weighted De-Mixing
BSD 2-Clause "Simplified" License
105 stars 30 forks source link

Lineage specific plot trouble shooting #220

Open Aishah86 opened 8 months ago

Aishah86 commented 8 months ago

Dear,

We are looking into how freyja performs plotting on specific lineage (--lineages) on the visualization. We have plot our samples and the visualization display majorities of the lineages were group under "other". But if we look at the aggregate data "other" were not listed. Can you clarify what is happening, or any other config that we need to do.

Thank you

Specific lineage SARS-CoV-2

Hereby i attached aggregate data for Feb, for your reference

aggregate data February 2023

-AIshah-

joshuailevy commented 7 months ago

Hi Aishah!

Given the outputs provided, it seems you may just have a bunch of low level variants. When running demix, I'd recommend setting --eps to a small values (say, 0.00001) and for plotting you can lower the --thresh value as well. It may also be helpful to use the --config option to specify lineage groupings/colors, otherwise you can run out of lineage colors quite quickly and you end up with a lot of "Other".

Best, Josh

Aishah86 commented 7 months ago

Do you mean the --thresh used during freyja plot? most of our data the plotting will be for data with < 0.5 abundance. if let say we put --thresh 0.1 , does it mean that lineages with >0.1 abundance will be plotted?

joshuailevy commented 6 months ago

Apologies- this seems to have evaded me. Yes, exactly. Under the hood this gets implemented via the prepLineageDict() function.