molBV

This is a brief instruction showing several options for installing and running the molBV algorithm (which calculated Nugent-like scores ranging from 1-10 using 16S rRNA data sequenced from cervicovaginal samples). See original publication for full details, validation and connection between molBV, cervicovaignal cytokines and cervix pre-cancer: https://www.nature.com/articles/s41467-021-27628-3

molBV Package

Installation

Download the source file molBV_1.0.tar.gz and run the following R commands:

install.packages("molBV_1.0.tar.gz", repos = NULL, type = "source")
browseVignettes("molBV")
??molBV

This will install the package and allow you to browse a sample tutorial using the above command. At the bottom of this page there is additional instructions for running the molBV algorithm manually (i.e. see "molBV (manual instructions)", which is more clunky, but you get to see all of the individual commands).

Please cite original paper if using molBV:

Usyk, M., Schlecht, N. F., Pickering, S., Williams, L., Sollecito, C. C., & Gradissimo, A. molBV reveals immune landscape of bacterial vaginosis and predicts human papillomavirus infection natural history. Nature Communications 2022; 13: 233. https://www.nature.com/articles/s41467-021-27628-3

Example Usage

created by Usyk, Mykhaylo: mykhaylo.usyk@einsteinmed.edu and Ling, Wodan: wling@fredhutch.org

Load the Package and Required Libraries

library(molBV)
library(phyloseq)
library(ggplot2)
library(ggpubr)

Load Example Data

data(exampleMicrobiomeData)
dim(exampleMicrobiomeData)
head(exampleMicrobiomeData)
range(exampleMicrobiomeData)

Calculate molBV Scores

map_with_molBVScores <- molBV(exampleMicrobiomeData, outputTable = NULL)
str(map_with_molBVScores)
range(map_with_molBVScores$molBV)

Use BIOM File for Input

exampleBiomFile <- system.file("extdata", "converted_otu.biom", package = "molBV")
map_with_molBVScores <- molBV(exampleBiomFile, outputTable = "map_with_molBV.txt")

Merge with Clinical Metadata

exampleMetaData <- system.file("extdata", "Mt_Sinai_sample_map.txt", package = "molBV")
clinical_map <- read.table(exampleMetaData, header = TRUE, sep = "\t", as.is = TRUE)
str(clinical_map)

full_set <- merge(x = map_with_molBVScores, y = clinical_map, by.x = "SID", by.y = "Cerv_SampleID")

Visualize molBV Scores compared to Clinical BV diagnosis

basic_boxes <- function(df_in=full_set, x_variable="Amsel_Diagnosis"){
	p <- ggplot(df_in, aes_string(x = x_variable, y = "molBV", fill = x_variable)) +
		geom_boxplot()+
		theme_minimal() + stat_compare_means() +
		ggtitle(label = paste("molBV across", x_variable))
	print(p)
}

basic_boxes()  # molBV across Amsel_Diagnosis
basic_boxes(x_variable = "Nugent_diagnosis")  # molBV across Nugent_diagnosis

molBV (manual instructions)

This is a brief instruction for the "rlm_molBV.R" script. The purpose of the script is to perform the calculations described in the manuscript tittled "molBV reveals immune landscape of bacterial vaginosis and predicts human papillomavirus infection natural history".

To run the script you should have R version 4.2.0 installed. Additionally if you will be importing data using the biom format, you will need to install the phyloseq package. The script also uses the ggplot2 and ggpubr packages to visualize the association of molBV with clinical variables and you will need to install these packages if you would like to generate the boxplots. Typical installation time for the packages should take under 10 minutes on MacBook Pro laptop with an i7 processor and 16 Gb of RAM.

Please cite original paper if using molBV: Usyk, M., Schlecht, N. F., Pickering, S., Williams, L., Sollecito, C. C., & Gradissimo, A. molBV reveals immune landscape of bacterial vaginosis and predicts human papillomavirus infection natural history. Nat Commun. 2022; 13: 233. https://www.nature.com/articles/s41467-021-27628-3