We previously developed a program in python3 that flags possible lab-associated variants in SARS-CoV-2 data (see https://github.com/lgozasht/COVID-19-Lab-Specific-Bias-Filter). This repository hosts our improved pipeline which employs similar methods as COVID-19-Lab-Specific-Bias-Filter but includes additional modifications.
We describe these modifications in detail in https://virological.org/t/issues-with-sars-cov-2-sequencing-data/473/12.
Briefly, our improved method requires an "unresolved" VCF as input, in which ambiguities remain unresolved (ambiguities can be a valuable signature of potentially errant alleles). It then performs two independent associations: one after resolving all ambiguous calls as the alternative allele (“alternative resolved”), and one where all ambiguous calls are resolved by parsimony (selecting the reference allele where there are equally parsimonious configurations: “parsimony resolved”). In addition, we added a “country-association” module, using the same methods we previously employed for detecting “lab-associated” variants. Our pipeline also associates specific alleles (including ambiguities) with specific labs and countries (i.e. if 99% of "W" allele calls at a given site are attributed to a particular lab, regardless if that lab contributes significantly to the total alternate allele count).
We also provide a separate program that identifies local positive linkage disequilibrium (LD) between variants. Local positive LD can be a signature of recurrent error.
Install from https://github.com/yatisht/strain_phylogenetics
Install from https://www.cog-genomics.org/plink/2.0/ and add to PATH
SARS-COV-2_COLLETIVE_ANALYSIS.py only requires a GISAID metadata file, unresolved VCF file, and corresponding newick tree as input. The pipeline will automatically "filter" the input VCF file for samples that also exist in the corresponding tree and will produce "filtered_unresolved.vcf" as a result. It will then produce "filtered_resolved_alt.vcf" after resolviong ambiguities as alternate alleles and "filtered_resolved_ref.vcf" after resolving ambiguities using parsimony. Note that the pipeline will not overwrite these files, so to rerun on a new dataset just delete them or remove them from your current working directory.
python3 SARS-COV-2_COLLETIVE_ANALYSIS.py [options] -m [Path to GISAID metadata file] -v [Path to VCF file] -tree [Path to newick tree] -o [Path to output directory]
GISAID metadata file, "unresolved" VCF and coorresponding newick tree
-min_parsimony I: Minimum parsimony (must be an integer) default = 4
-dependencies: Check for dependencies
-threads I: Number of threads to use per association
final_table.tsv:
| Column | Description |
|---|---|
| Reference | reference genome |
| Start | site start |
| Stop | site stop |
| Snp | snp |
| Alt Resolved Parsimony | "alternative resolved" parsimony score |
| MAC | "alternative resolved" minor allele count |
| MAF | "alternative resolved" minor allele frequency |
| Primer Overlap | ARTIC primer overlapping site (NA if none) |
| Primer Vicinity | ARTIC primer within 10bp of site (NA if none) |
| Country | "alternative resolved" country associated with variant (NA if none) |
| Country Association | "alternative resolved" percent of minor alleles attributed to that country |
| Lab | "alternative resolved" lab associated with variant (NA if none) |
| Lab Association | "alternative resolved" percent of minor alleles attributed to that lab |
| Ref Resolved Parsimony | "parsimony resolved" parsimony score |
| MAC | "parsimony resolved" minor allele count |
| MAF | "parsimony resolved" minor allele frequency |
| Country | "parsimony resolved" country associated with variant (NA if none) |
| Country Association | "parsimony resolved" percent of minor alleles attributed to that country |
| Lab | "parsimony resolved" lab associated with variant (NA if none) |
| Lab Association | "parsimony resolved" percent of minor alleles attributed to that lab |
| Specific Alt Associations | Specific alt associations populate the remaining columns (depending on how many alternate alleles exist at a given site) |
samples_in_latest_tree.txt: List of samples that exist both in the provided VCF and the corresponding newick tree.
filtered_unresolved.vcf: Unresolved VCF containing only samples that exist both in the provided VCF and the corresponding newick tree.
filtered_unresolved.txt: Parsimony file corresponding to the unresolved VCF.
filtered_resolved_alt.vcf: "Alternate resolved" VCF
filtered_resolved_alt.txt: "Alternate resolved" parsimony file
filtered_resolved_ref.vcf: "Parsimony resolved" VCF
filtered_resolved_ref.txt: "Parsimony resolved" parsimony file
Calculates local R^2 to identify linked variants.
python3 local_LD.py [options] -v [Path to VCF file] -table [Path to "final_table.tsv" (produced by SARS-COV-2_COLLETIVE_ANALYSIS.py)] -o [Path to output directory]
VCF and "final_table.tsv" (produced by SARS-COV-2_COLLETIVE_ANALYSIS.py)
-range I: Max range in which LD will be estimated for a given site (Default = 10bp)
-dependencies: Check for dependencies
linked_sites.tsv:
| Column | Description |
|---|---|
| snp | snp |
| pos | site position |
| linked site|r^2 | linked site | corresponding r^2 value |
-
Landen Gozashti, Conor Walker, Robert Lanfear, Nick Goldman, Nicola De Maio and Russell Corbett-Detig, "Issues with SARS-Cov-2 sequencing data: Updated analysis with data from 4 March 2021", Virological post 2021.
-
Yatish Turakhia, Nicola De Maio, Bryan Thornlow, Landen Gozashti, Robert Lanfear, Conor R. Walker, Angie S. Hinrichs, Jason D. Fernandes, Rui Borges, Greg Slodkowicz, Lukas Weilguny, David Haussler, Nick Goldman and Russell Corbett-Detig, "Stability of SARS-CoV-2 Phylogenies", PLOS Genetics 2020 (https://doi.org/10.1371/journal.pgen.1009175).