New preprint: processing 3.9 million SARS-CoV-2 samples to make a consistent phylogenetic tree

Martin Hunt, Zam Iqbal and lots of others have written an epic preprint where they describe their variant caller, viridian, that was written expressly for producing a consensus genome for a virus using tiled amplicon sequencing. We deployed viridian into our cloud-based sequencing platform during the pandemic and several of us are co-authors on the preprint.

Rather than just write a straightforward methods paper, they choose to take it a few steps further and download all the publicly available SARS-CoV-2 FASTQ files from the ENA (which was about 3.9 million) and process them with viridian and then build a phylogenetic tree that is missing some of the artefacts that plague the trees if you simply take FASTA files that have been deposited using a range of genetic pipelines.