Download


MLP can be downloaded as an R package from GitHub and used locally as well.

GitHub: https://github.com/grp-bork/microbial_load_predictor



Requirements

  • R 4.3.1+
  • vegan
  • tidyverse


How to start

  • On Terminal:
    • git clone https://github.com/grp-bork/microbial_load_predictor

  • Basic usage on R:
    • devtools::install() 
library("MLP") 

# Predicting microbial load
load <- MLP(input, "motus2", "load") 

# Transforming relative microbiome profile (RMP) to quantitative microbiome profile (QMP)
qmp <- MLP(input, "motus2", "qmp") 

# Quantitative (absolute) abundance = relative abundance * predicted microbial load


Example code using test data

# read input file (mOTUs v2.5) 
input <- read.delim("test_data/Franzosa_2018_IBD.motus25.tsv", header = T, row.names = 1, check.names = F) 

# transpose the data 
input <- data.frame(t(input), check.names = F) 

# predict microbial loads 
load <- MLP(input, "motus2", "load") 

# transform relative microbiome profile (RMP) to quantitative microbiome profile (QMP) 
qmp <- MLP(input, "motus2", "qmp") 

# plot predicted microbial loads
md <- read.delim("test_data/Franzosa_2018_IBD.metadata.tsv", header = T, row.names = 1, check.names = F) 
df <- data.frame(md, load = load$load) 
ggplot(df, aes(x = Disease, y = log10(load), fill = Disease)) + 
    theme_bw() + 
    geom_boxplot()


The test data comes from Franzosa EA et al., 2018 including Crohn's disease and ulcerative colitis patients as well as control individuals.

Gut microbiome structure and metabolic activity in inflammatory bowel disease

Eric A Franzosa, Alexandra Sirota-Madi, Julian Avila-Pacheco, Nadine Fornelos, Henry J Haiser, Stefan Reinker, Tommi Vatanen, A Brantley Hall, Himel Mallick, Lauren J McIver, Jenny S Sauk, Robin G Wilson, Betsy W Stevens, Justin M Scott, Kerry Pierce, Amy A Deik, Kevin Bullock, Floris Imhann, Jeffrey A Porter, Alexandra Zhernakova, Jingyuan Fu, Rinse K Weersma, Cisca Wijmenga, Clary B Clish, Hera Vlamakis, Curtis Huttenhower, Ramnik J Xavier

Nat Microbiol. 2019 Feb;4(2):293-305. doi: 10.1038/s41564-018-0306-4.