prewas: Preprocess SNPs before bGWAS

In prewas: Data Pre-Processing for Bacterial Genome-Wide Association Studies

Description

prewas is a tool to standardize the pre-processing of your genomic data before performing a bacterial genome-wide association study (bGWAS). prewas creates a variant matrix (where each row is a variant, each column is a sample, and the entries are presence - 1 - or absence - 0 - of the variant) that can be used as input for bGWAS tools. When creating the binary variant matrix, prewas can perform 3 pre-processing steps including: dealing with multiallelic SNPs, (optional) dealing with SNPs in overlapping genes, and choosing a reference allele. prewas can output matrices for use with both SNP-based bGWAS and gene-based bGWAS. prewas can also provide gene matrices for variants with specific SnpEff annotations (Cingolani et al. 2012).

Usage

prewas(
  dna,
  tree = NULL,
  outgroup = NULL,
  gff = NULL,
  anc = TRUE,
  snpeff_grouping = NULL,
  grp_nonref = FALSE
)

Arguments

dna	'Character' or 'vcfR'. Required input. Path to VCF4.1 file or 'vcfR' object.
tree	'NULL', 'character', or 'phylo'. Optional input. Ignored if 'NULL'. If 'character' it should be a path to a .tree file. Defaults to 'NULL'.
outgroup	'NULL' or 'character'. Optional input. If 'character' it should be either a string naming the outgroup in the tree or a path to a file containing only the outgroup name. Ignored if 'NULL'. Defaults to 'NULL'.
gff	'NULL', 'character', 'matrix', or 'data.frame'. Optional input. If 'NULL' it is ignored. If 'character' it should be a path to a GFF3 file. If a 'matrix' or 'data.frame' it should be the GFF information stored in 9 columns with the genes as rows. Defaults to 'NULL'.
anc	'Logical'. Optional input. When 'TRUE' prewas performs ancestral reconstruction. When 'FALSE' prewas calculates the major allele. Defaults to 'TRUE'.
snpeff_grouping	'NULL', 'character'. Optional input. Only used when a SnpEff annotated multivcf is inputted. Use when you want to group SNPs by gene and SnpEff impact. If 'NULL' no custom-grouped gene matrix will be generated. Options for input are a vector combination of 'HIGH', 'MODERATE', LOW', 'MODIFER'. Must write the impact combinations in all caps (e.g. c('HIGH', 'MODERATE')). Defaults to 'NULL'.
grp_nonref	'Logical'. Optional input. When 'TRUE' prewas collapses all non-reference alleles for multi-allelic sites. When 'FALSE' prewas keeps multi-allelic sites separate. Defaults to 'FALSE'.

Value

A list with the following items:

allele_mat

'matrix'. An allele matrix, created from the vcf where each multiallelic site will be on its own line. The rowname will be the position of the variant in the vcf file. If the position is triallelic, there will be two rows containing the same information. The rows will be labeled "pos" and "pos.1". If the position is quadallelic, there will be three rows containing the same information. The rows will be labeled "pos", "pos.1", and "pos.2"

bin_mat

'matrix'. A binary matrix, where 0 is the reference allele and 1 indicates a variant. The dimensions may not match the 'allele_mat' if the gff file is provided, because SNPs in overlapping genes are represented on multiple lines in 'bin_mat'; in that case both position and locus tag name are provided in the rowname.

ar_results

'data.frame'. This data.frame records the alleles used as the reference alleles. Rows correspond to variant loci. If 'anc = TRUE' the data.frame has two columns which contain the ancestrally reconstructed allele and the probability of the reconstruction. If 'anc = FALSE' there is only one column which contains the major allele.

dup

‘integer'. A vector of integers. It’s an index that identifies duplicated rows. If the index is unique (appears once), that means it is not a multiallelic site. If the index appears more than once, that means the row was replicated 'x' times, where 'x' is the number of alternative alleles. Note: the multiple indices indicates multiallelic site splits, not overlapping genes splits.

gene_mat

'NULL' or 'matrix' or 'list'. 'NULL' if no gene information provided ('gff = NULL' and no SnpEff annotation provided in VCF). If gene information is provided by a gff then a gene matrix is generated where each row is a gene and each column is a sample. If gene information is provided by a SnpEff annotated vcf then a list of up to six gene matrices are returned. The first matrix, 'gene_mat_all' is a gene matrix for all variants.'gene_mat_modifier' is a gene matrix for only variants annoted as MODIFIER by SnpEff. Similarly there is a 'gene_mat_low', 'gene_mat_moderate', and 'gene_mat_high.' If the user asks for a combination SnpEff annotations the final 'gene_mat_custom' will contain that matrix.

tree

'NULL' or 'phylo'. If 'anc = FALSE' no tree is use or generated and the function returns 'NULL'. If 'anc = TRUE' and the user provides a tree but no outgroup: the function returns the tree after midpoint rooting. If 'anc = TRUE' and the user provides both a tree and an outgroup: the function returns a tree rooted on the outgroup and the outgroup is removed from the tree. If the user does not provide a tree and 'anc = TRUE' the function returns the midpoint rooted tree generated.

Examples

vcf = prewas::vcf
gff = prewas::gff
tree = prewas::tree
outgroup = prewas::outgroup
output <- prewas(dna = vcf,
                 tree = tree,
                 outgroup = outgroup,
                 gff = gff,
                 anc = FALSE ,
                 grp_nonref = FALSE)

prewas documentation built on April 2, 2021, 5:06 p.m.