roblanf/sarscov2phylo: 18-10-20

Citation and reuse

Please cite this release as:

Lanfear, Rob (2020). A global phylogeny of SARS-CoV-2 sequences from GISAID. Zenodo DOI: 10.5281/zenodo.3958883

You can visit that DOI here:

If you publish papers that use this tree you must still follow the GISAID data sharing and attribution rules.

Details

The trees in this release were generated with the following command line:

bash global_tree_gisaid_start_tree.sh -i [gisaid.fasta] -p [previous_iteration] -t 250

• [gisaid.fasta] is the fasta file of high coverage and complete raw sequences from GISAID up to and including the date in the title of the release, determined by the 'submission date' filter on a GISAID data feed

• [previous_iteration] is the filepath of the previous release, this is used to provide the excluded_sequences.tsv and ft_SH.tree files as the starting points of the current iteration.

Filtering statistics

sequences downloaded from GISAID
108202
//
alignment stats of global alignment
Alignment number:    1
Format:              aligned FASTA
Number of sequences: 106464
Alignment length:    29903
Total # residues:    3173059520
Smallest:            29030
Largest:             29903
Average length:      29804.1
Average identity:    100%
//
alignment stats of global alignment after masking sites
Alignment number:    1
Format:              aligned FASTA
Number of sequences: 106464
Alignment length:    29903
Total # residues:    3158473719
Smallest:            28961
Largest:             29675
Average length:      29667.1
Average identity:    100%
//
alignment stats after filtering out short/ambiguous sequences
Alignment number:    1
Format:              aligned FASTA
Number of sequences: 106273
Alignment length:    29903
Total # residues:    3152816993
Smallest:            28961
Largest:             29675
Average length:      29667.1
Average identity:    100%
//
alignment stats of global alignment after trimming sites that are >50% gaps
Alignment number:    1
Format:              aligned FASTA
Number of sequences: 106273
Alignment length:    29646
Total # residues:    3144363035
Smallest:            28437
Largest:             29646
Average length:      29587.6
Average identity:    100%
//
After filtering sequences with TreeShrink
Type:   Phylogram
#nodes: 186198
#leaves:    106272
#dichotomies:   75928
#leaf labels:   106272
#inner labels:  71407
Number of new sequences added this iteration
340 alignment_names_new.txt

Notable changes to the scripts in this release

• Some more issues occurred with sequence names, so I have shifted to using exclusively just the EPI-IDs for sequence names in the alignments, trees, and scripts. This should avoid any future issues with sequence names.

If you have a GISAID data stream, this means you can get the data as follows:

curl -u "${GISAID_LOGIN}:${GISAID_PASSWORD}" ${URL} 2>/dev/null \
    | xz -d -T22 \
    | jq -r 'select(.covv_subm_date <= "'"$DATE"'" and .is_complete == true and .is_high_coverage == true) | [">" + .covv_accession_id + "\n" + .sequence] | join("\n")' \
    | xz -c -T22 > gisaid-${DATE}.fasta.xz

Notable aspects of the trees

• Names have changed to include only the EPI-IDs. This has the side-benefit of roughly halving the size of the tree file.