Pandemic-scale phylogenetics

Abstract

Phylogenetics has been central to the genomic surveillance, epidemiology and contact tracing efforts during the COVD-19 pandemic. But the massive scale of genomic sequencing has rendered the pre-pandemic tools inadequate for comprehensive phylogenetic analyses. Here, we discuss the phylogenetic package that we developed to address the needs imposed by this pandemic. The package incorporates several pandemic-specific optimization and parallelization techniques and comprises four programs: UShER, matOptimize, RIPPLES and matUtils. Using high-performance computing, UShER and matOptimize maintain and refine daily a massive mutation-annotated phylogenetic tree consisting of all SARS-CoV-2 sequences available in online repositories. With UShER and RIPPLES, individual labs -- even with modest compute resources -- incorporate newly-sequenced SARS-CoV-2 genomes on this phylogeny and discover evidence for recombination in real-time. With matUtils, they rapidly query and visualize massive SARS-CoV-2 phylogenies. These tools have empowered scientists worldwide to study the SARS-CoV-2 evolution and transmission at an unprecedented scale, resolution and speed.

Identifying SARS-CoV-2 regional introductions and transmission clusters in real time

“…Using our method, we traced transmission clusters in 102 countries from across the world (Figure 2A) using the global parsimony phylogenetic tree, built from 5,563,847 available sequences on GISAID 21 , GenBank 19 , and COG-UK 25 on 11-28-2021 (see Methods). Cluster size is highly skewed (Figure 2C), with approximately 20% of distinct regional clusters containing 89% of samples.…”

“…At UCSC we maintain a large phylogeny of all GISAID 21 , GenBank 19 , and COG-UK 25 sequences using the script https://github.com/ucscGenomeBrowser/kent/blob/master/src/hg/utils/otto/sarscov2phylo/update Public.sh and the UShER online phylogenetics suite 12,21 . Updates are performed daily by obtaining all newly uploaded sequences from each database and placing them on the previous day’s global phylogenetic tree with UShER (see McBroome et al) 13 .Starting with our phylogeny updated on 11-28-2021, we pruned all samples with long branch lengths and path lengths using the matUtils parameters --max-branch-length 45 and --max-path-length 100 and performed a round of optimization with an SPR radius of 8.…”

“…Some have made progress by adopting generalized statistical methods built for large data such as random forest regression 14 , but scalable phylogenetic solutions need to be developed. While our group, among others, has laid the groundwork for pandemic-scale phylogenetics 3,12,13,20,23,25 , much remains to be done to translate evolutionary inferences to public health inference and response.…”

Variational Phylodynamic Inference Using Pandemic-scale Data

“…A phylogenetic representation of this database is believed to be the largest ever constructed ( Turakhia, Thornlow, Hinrichs, De Maio, et al 2021 ). Existing phylogenetic methods, which were developed and tested on datasets orders of magnitude smaller, are inadequate for pandemic-scale analysis, resulting in missed opportunities to improve our surveillance and response capabilities ( Hodcroft et al 2021 ; Morel et al 2021 ; Ye et al 2021 ).…”

Variational phylodynamic inference using pandemic-scale data

“…A phylogenetic representation of this database is believed to be the largest ever constructed (Turakhia et al, 2021a). Existing phylogenetic methods, which were developed and tested on datasets orders of magnitude smaller, are inadequate for pandemic-scale analysis, resulting in missed opportunities to improve our surveillance and response capabilities (Hodcroft et al, 2021; Ye et al, 2021; Morel et al, 2021).…”