Comparing fitness dynamics across SARS-CoV-2, influenza H3 and influenza H1

Provision metadata locally

mkdir data
cd data

For SARS-CoV-2

aws s3 cp s3://nextstrain-ncov-private/metadata.tsv.zst sarscov2_metadata.tsv.zst
zstd -c -d sarscov2_metadata.tsv.zst \
   | tsv-select -H -f strain,date,country,clade_nextstrain,Nextclade_pango,QC_overall_status \
   | zstd -c > sarscov2_subset_metadata.tsv.zst

and move to fitness-dynamics/data/.

For H3N2, clone https://github.com/blab/flu-geo-fitness and then run

nextstrain build . data/h3n2/metadata_with_nextclade.tsv
cd data/h3n2/
sed -i -e 's/\tseqName\t/\tstrain\t/' metadata_with_nextclade.tsv
sed -i -e 's/\tUsa\t/\tUSA\t/g' metadata_with_nextclade.tsv
sed -i -e '1s/$/\tinclusion/; 2,$s/$/\tglobal/' metadata_with_nextclade.tsv
tsv-select -H -f strain,date,region,country,inclusion,subclade,qc.overallStatus metadata_with_nextclade.tsv > metadata_selected.tsv
tsv-filter -H --str-ne subclade:unassigned --str-gt date:2000-01-01 metadata_selected.tsv > metadata_filtered.tsv
zstd -c metadata_filtered.tsv > h3n2_subset_metadata.tsv.zst

and move to fitness-dynamics/data/.

Workflow

Once metadata is provisioned locally, run the entire workflow with

nextstrain build . all_mlr_estimates

Sequence counts

Data for the project consists of daily sequence counts of clades of SARS-CoV-2, influenza H3 and influenza H1. Sequence counts are provisioned to the sequence-counts/ directory. From top-level directory run

nextstrain build . all_sequence_counts

to produce the sequence counts files

sequence-counts/sarscov2_clades_2020/prepared_seq_counts.tsv

Currently, clade counts are provisioned for just the USA.

MLR estimates

Run MLR models using evofr package. Run the model with

nextstrain build . all_mlr_estimates

to produce the MLR output JSON files

mlr-estimates/sarscov2_clades_2020/mlr_results.json

Scaffolded fitnesses

Fitnesses within each timepoint are measured relative to an arbitrary pivot variant. The Mathematica notebook fitness-flux-analysis/fitness-flux.nb takes mlr_results.json across timepoints and combines into a single scaffolded-fitness/sarscov2_clades_scaffolded_fitness.tsv. This notebook needs to be run separately for virus sarscov2 classification clades, virus sarscov2 classification lineages and virus h3n2 classification clades.