,
Steven Schiff [aut],
Joseph N Paulson [aut]| Title: | mbQTL: A package for SNP-Taxa mGWAS analysis |
|---|---|
| Description: | mbQTL is a statistical R package for simultaneous 16srRNA,16srDNA (microbial) and variant, SNP, SNV (host) relationship, correlation, regression studies. We apply linear, logistic and correlation based statistics to identify the relationships of taxa, genus, species and variant, SNP, SNV in the infected host. We produce various statistical significance measures such as P values, FDR, BC and probability estimation to show significance of these relationships. Further we provide various visualization function for ease and clarification of the results of these analysis. The package is compatible with dataframe, MRexperiment and text formats. |
| Authors: | Mercedeh Movassagh [aut, cre] ,
Steven Schiff [aut],
Joseph N Paulson [aut] |
| Maintainer: | Mercedeh Movassagh <[email protected]> |
| License: | MIT + file LICENSE |
| Version: | 1.7.0 |
| Built: | 2024-10-30 09:16:59 UTC |
| Source: | https://github.com/bioc/mbQTL |
mbQTL is a package for microbial QTL/GWAS AnalysisThis package provides statistical methods for identifying significant relationships between
microbial/taxa and genetic SNP signatures. We use three models 1) linear regression between
all taxa-snp. Main function is linearTaxaSnp(). 2) Correlation analysis between taxa-snp
across all taxa and snps. Main function is taxaSnpCor() and 3) Logistic regression analysis
between each taxa and each snp simultaneously or for a specific cases. Main function is
logRegSnpsTaxa().
Maintainer: Mercedeh Movassagh [email protected] (ORCID)
Authors:
Steven Schiff
Joseph N Paulson
The package vignette can be accessed with vignette("mbQTL").
allToAllProduct creates a dataframe of snp and taxa correlationsThis internal function takes the original snp dataframe and returns a long parsed snp dataframe
allToAllProduct(SnpFile, microbeAbund, rsID = NULL)allToAllProduct(SnpFile, microbeAbund, rsID = NULL)
SnpFile |
the snp file (rownames is sample number and colnames is the snps) |
microbeAbund |
he taxa abundance dataframe (rownames sample names and colnames taxa Genus/species/family) |
rsID |
Default is NULL and will run across the who dataset unless specific rsID/SNP/chr_region is specified |
A dataframe of correlations between snps and taxa
data(microbeAbund) data(SnpFile) x <- allToAllProduct(SnpFile, microbeAbund, "chr1.171282963_T")data(microbeAbund) data(SnpFile) x <- allToAllProduct(SnpFile, microbeAbund, "chr1.171282963_T")
binarizeMicrobe binarizes microbe abundace file based on user's cutoffThis function creates a dataframe output produces a formatted dataframe prepared.
binarizeMicrobe(microbeAbund, cutoff = NULL, selectmicrobe = NULL)binarizeMicrobe(microbeAbund, cutoff = NULL, selectmicrobe = NULL)
microbeAbund |
the taxa abundance dataframe (rownames sample names and colnames taxa Genus/species/family) |
cutoff |
cutoff at which the user chose to call taxa positive or negative across samples (should be a numeric value for normalized or count values). |
selectmicrobe |
default is and all taxa are considered at the same time, if the user is interested in a specific pathogen use name of the pathogen for example "Haemophilus". |
A data frame of microbial abundance.
coringTaxa creates correlation dataframe for taxaThis function creates an output correlation data frame for all microbial taxa (or other organisms such as viral or parasitic taxa)
coringTaxa(microbeAbund)coringTaxa(microbeAbund)
microbeAbund |
the taxa abundance dataframe (rownames sample names and colnames taxa Genus/species/family) |
A data frame of correlations between taxa
data(microbeAbund) x <- coringTaxa(microbeAbund)data(microbeAbund) x <- coringTaxa(microbeAbund)
mbQTL "CovFile"The "CovFile" is the covariate file for linear regression option of taxa and snp association. The covariance file is generated randomly by assigning sex and site of collection to the samples.) rownames are covariate and colnames samples.
histPvalueLm a histogram of Taxa and SNP linear regression analysis. This function creates a histogram object of all SNPs with all taxa Linear regression analysis p values.
histPvalueLm(LinearAnalysisTaxaSNP)histPvalueLm(LinearAnalysisTaxaSNP)
LinearAnalysisTaxaSNP |
the data frame result created from the |
A histogram object of p values observed from taxa and SNP Linear Regression analysis.
data(microbeAbund) data(microbeAbund) data(SnpFile) data(CovFile) LinearAnalysisTaxaSNPFile <- linearTaxaSnp(microbeAbund, SnpFile, Covariate = CovFile) x <- histPvalueLm(LinearAnalysisTaxaSNPFile)data(microbeAbund) data(microbeAbund) data(SnpFile) data(CovFile) LinearAnalysisTaxaSNPFile <- linearTaxaSnp(microbeAbund, SnpFile, Covariate = CovFile) x <- histPvalueLm(LinearAnalysisTaxaSNPFile)
This function creates a dataframe output from the results all snps with all taxa linear regression analysis of all snps in the dataset. The result is a dataframe with P values and FDRs of all regressions. MatrixeQTL core functions are utilized to achieve this. Note the main functions used are Matrix_eQTL_engine() assuming linear regression with or without a covariate file.
linearTaxaSnp(microbeAbund, SnpFile, Covariate = NULL)linearTaxaSnp(microbeAbund, SnpFile, Covariate = NULL)
microbeAbund |
the taxa abundance dataframe (rownames sample names and colnames taxa Genus/species/family) |
SnpFile |
the snp dataframe (values 0,1,2 indicating zygosity), rownames sample names and colnames snp names. |
Covariate |
default is NULL, hence assumed non-existent. If covariates are available they need to be formatted in the CovFile format, that is colnames are sample numbers matching samples in the microbe abundance and snp file and row names are the co-variates names (such as sex, disease etc). |
A data frame which is a result of Linear Regression of all snp, taxa relationships, with P values and P value corrected values.
data(microbeAbund) data(SnpFile) data(CovFile) x <- linearTaxaSnp(microbeAbund, SnpFile, Covariate = CovFile)data(microbeAbund) data(SnpFile) data(CovFile) x <- linearTaxaSnp(microbeAbund, SnpFile, Covariate = CovFile)
This function creates a dataframe output produces a formatted dataframe prepared.
logitPlotSnpTaxa( microbeAbund, SnpFile, selectmicrobe = NULL, rsID, ref = NULL, alt = NULL, het = NULL, color = NULL, cutoff = NULL )logitPlotSnpTaxa( microbeAbund, SnpFile, selectmicrobe = NULL, rsID, ref = NULL, alt = NULL, het = NULL, color = NULL, cutoff = NULL )
microbeAbund |
original microbe abundance file (colnames microbe, rownames= sample IDs) |
SnpFile |
original snp file with (0,1,2 values for ref, het, alt genotypes), colnames SNP names, rownames, sample IDs. |
selectmicrobe |
name of the microbe of interest (for example individual significant microbes associate with a snp). |
rsID |
name of the snp of interest (for example individual significant snps associated with a microbe) |
ref |
the name of reference genotype for example "GG" |
alt |
the name of snp (variant) genotype for example "AA" |
het |
the name of hetrozygote genotype for example "GA" |
color |
the default is NULL and the color is set to c("#ffaa1e", "#87365b"). |
cutoff |
cutoff at which we call microbe present or absent |
A bar ggplot comparing the counts of ref vs alt vs het genotype
data(microbeAbund) data(SnpFile) x <- logitPlotSnpTaxa(microbeAbund, SnpFile, selectmicrobe = "Neisseria", rsID = "chr2.241072116_A", ref = NULL, alt = NULL, het = NULL, color = NULL, cutoff = NULL )data(microbeAbund) data(SnpFile) x <- logitPlotSnpTaxa(microbeAbund, SnpFile, selectmicrobe = "Neisseria", rsID = "chr2.241072116_A", ref = NULL, alt = NULL, het = NULL, color = NULL, cutoff = NULL )
logRegSnpsTaxa Performs logistic regression analysis between taxa and SNPs and returns concordance statisticsThis function creates a dataframe output from the results of either a unique taxa and all snps or all taxa and all snps in the dataset. The result is a dataframe with P values and FDRs of all regressions.
logRegSnpsTaxa(microbeAbund, SnpFile, cutoff = NULL, selectmicrobe = NULL)logRegSnpsTaxa(microbeAbund, SnpFile, cutoff = NULL, selectmicrobe = NULL)
microbeAbund |
the taxa abundance dataframe (rownames sample names and colnames taxa Genus/species/family) |
SnpFile |
the snp dataframe (values 0,1,2 indicating zygosity), rownames sample names and colnames snp names. |
cutoff |
default is NULL, hence anything above cutoff is considered positive, the cutoff at which the specific or all taxa are considered positive for the pathogen (1 indicates positive and 0 negative). |
selectmicrobe |
default is and all taxa are considered at the same time, if the user is interested in a specific pathogen use name of the pathogen for example "Haemophilus". |
A data frame which is a result of Logistic regression products of individual snp, taxa relationships, with P values and P value corrected values (FDR, Bonferroni).
data(microbeAbund) data(SnpFile) x <- logRegSnpsTaxa(microbeAbund, SnpFile, selectmicrobe = c("Haemophilus"))data(microbeAbund) data(SnpFile) x <- logRegSnpsTaxa(microbeAbund, SnpFile, selectmicrobe = c("Haemophilus"))
mbQtlCorHeatmap for making heatmap for snp, taxa rho valuesThis function produces a log heatmap +1 of the correlation rho values across snp, taxa datasets
mbQtlCorHeatmap(final_var_long, labels_col = NULL, ...)mbQtlCorHeatmap(final_var_long, labels_col = NULL, ...)
final_var_long |
the long data frame of rho values created by the |
labels_col |
set to NULL ass default if TRUE, labels will appear on the heatmap. |
... |
all other parameters for pheatmap. |
A data frame of correlations between taxa
data(microbeAbund) data(SnpFile) for_all_rsids <- allToAllProduct(SnpFile, microbeAbund) correlationMicrobes <- coringTaxa(microbeAbund) taxaSnpCor(for_all_rsids, correlationMicrobes) final_var_long <- taxaSnpCor(for_all_rsids, correlationMicrobes, probs = c(0.0001, 0.9999)) x <- mbQtlCorHeatmap(final_var_long)data(microbeAbund) data(SnpFile) for_all_rsids <- allToAllProduct(SnpFile, microbeAbund) correlationMicrobes <- coringTaxa(microbeAbund) taxaSnpCor(for_all_rsids, correlationMicrobes) final_var_long <- taxaSnpCor(for_all_rsids, correlationMicrobes, probs = c(0.0001, 0.9999)) x <- mbQtlCorHeatmap(final_var_long)
mbQTL "metagenomeSeqObj"
"MetagenomSeqObj" is an MRexperiment object format of the "microbeAbund" file.mbQTL "metagenomeSeqObj"
"MetagenomSeqObj" is an MRexperiment object format of the "microbeAbund" file.
metagenomeSeqToMbqtl Converts metagenomeSeq obj to compatible taxa dataframe
This function takes and MRexperiement class object transforms it and makes the result dataframe compatible
with mbQTL taxa input file
metagenomeSeqToMbqtl(meta_glom, norm, log, aggregate_taxa = NULL)metagenomeSeqToMbqtl(meta_glom, norm, log, aggregate_taxa = NULL)
meta_glom |
|
norm |
A logical indicating whether or not to return normalized counts. |
log |
TRUE/FALSE whether or not to log2 transform scale. |
aggregate_taxa |
it is recommended that the normalization occurs at taxa level (default NULL) however, if the user chooses to aggregate on the phyla/family/Genus or Species level before normalization they have the option. |
A data frame of normalized/not normalized counts compatible with mbQTL.
data(metagenomeSeqObj) x <- metagenomeSeqToMbqtl(metagenomeSeqObj, norm = TRUE, log = TRUE, aggregate_taxa = NULL)data(metagenomeSeqObj) x <- metagenomeSeqToMbqtl(metagenomeSeqObj, norm = TRUE, log = TRUE, aggregate_taxa = NULL)
mbQTL "microbiomeAbund" FileThis is the microbial Abudnance file generated from 16s it is either this file or
the "metaGenomeSeqObj".The "microbiomeAbund" file is a randomly generated file
format for total microbe presence (number of reads)(parasite/viral transcripts)
for specific species.This could be generated from select taxa results from
biom() or MRexperiment objects as long as the samples are colnames and taxa
are rownames.
prepareCorData prpares and joins snp-taxa and taxa-taxa correlation file.This function creates a dataframe output produces a formatted dataframe prepared.
prepareCorData(microbeAbund, SnpFile, cutoff = NULL, selectmicrobe = NULL)prepareCorData(microbeAbund, SnpFile, cutoff = NULL, selectmicrobe = NULL)
microbeAbund |
the taxa abundance dataframe (rownames sample names and colnames taxa Genus/species/family) |
SnpFile |
the snp dataframe (values 0,1,2 indicating zygosity), rownames sample names and colnames snp names. |
cutoff |
default is NULL, hence anything above cutoff is considered positive, the cutoff at which the specific or all taxa are considered positive for the pathogen (1 indicates positive and 0 negative). |
selectmicrobe |
default is and all taxa are considered at the same time, if the user is interested in a specific pathogen use name of the pathogen for example "Haemophilus". |
A data frame which is a result of Logistic regression products of individual snp, taxa relationships, with P values and P value corrected values.
qqPlotLm creates QQ-Plot of all SNPs with all taxa Linear regression analysis This function creates QQ-Plot object of all SNPs with all taxa Linear regression analysis of expected versus observed P values
qqPlotLm(microbeAbund, SnpFile, Covariate = NULL)qqPlotLm(microbeAbund, SnpFile, Covariate = NULL)
microbeAbund |
the taxa abundance dataframe (rownames sample names and colnames taxa Genus/species/family) |
SnpFile |
the snp dataframe (values 0,1,2 indicating zygosity), rownames sample names and colnames snp names. |
Covariate |
default is NULL, hence assumed non-existent. If covariates are available they need to be formatted in the CovFile format, that is colnames are sample numbers matching samples in the microbe abundance and snp file and row names are the covariates names (such as sex, disease etc). |
A QQplot object of expected versus obsesrved taxa and SNP Linear Regression analysis
data(microbeAbund) data(SnpFile) data(CovFile) x <- qqPlotLm(microbeAbund, SnpFile, Covariate = CovFile)data(microbeAbund) data(SnpFile) data(CovFile) x <- qqPlotLm(microbeAbund, SnpFile, Covariate = CovFile)
RegSnp creates a dataframe of parsed long snp filesThis internal function takes the orginal snp dataframe and retruns a long parsed snp dataframe
RegSnp(SnpFile, microbeAbund)RegSnp(SnpFile, microbeAbund)
SnpFile |
the snp file (rownames is sample number and colnames is the snps) |
microbeAbund |
the microbial abundance file (rownames is sample number and colnames is the microbe) |
A long parsed datframe of snps
mbQTL "SnpFile"The "SnpFile" is randomly generated from GATK (Van der Auwera GA & O'Connor BD. (2020). Genomics in the Cloud: Using Docker, GATK, and WDL in Terra (1st Edition). O'Reilly Media) snp calls followed by plink (Purcell S, et al. (2007) PLINK: a toolset for whole-genome association and population-based linkage analysis. American Journal of Human Genetics) processing and it needs to be in (0,1,2) format representing the zygosity of the snps.
taxaSnpCor for estimation of the rho value between snp, taxa correlations across datasetsThis function produces a log heatmap +1 of the correlation rho values across snp, taxa dataframe.
taxaSnpCor(for_all_rsids, correlationMicrobes, probs = NULL)taxaSnpCor(for_all_rsids, correlationMicrobes, probs = NULL)
for_all_rsids |
A dataframe result of correlation analysis between the snps and taxa dataframe,
an output of |
correlationMicrobes |
A dataframe of correlation between |
probs |
Default is NULL if other that all rho values are wanted the value can be subseted using c( |
A data frame of correlations between taxa
data(microbeAbund) data(SnpFile) for_all_rsids <- allToAllProduct(SnpFile, microbeAbund) correlationMicrobes <- coringTaxa(microbeAbund) x <- taxaSnpCor(for_all_rsids, correlationMicrobes)data(microbeAbund) data(SnpFile) for_all_rsids <- allToAllProduct(SnpFile, microbeAbund) correlationMicrobes <- coringTaxa(microbeAbund) x <- taxaSnpCor(for_all_rsids, correlationMicrobes)