In Duuudude/lilikoi2: Metabolomics Personalized Pathway Analysis Tool

knitr::opts_chunk$set(echo = TRUE)

Lilikoi2: a Deep-learning enabled, personalized pathway-based R package for diagnosis and prognosis predictions using metabolomics data

Previously we developed Lilikoi, a personalized pathway-based method to classify diseases using metabolomics data. Given the new trends of computation in the metabolomics field, here we report the next version of Lilikoi as a significant upgrade. The new Lilikoi v2 R package has implemented a deep-learning method for classification, in addition to popular machine learning methods. It also has several new modules, including the most significant addition of prognosis prediction, implemented by Cox-PH model and the deep-learning based Cox-nnet model. Additionally, Lilikoi v2 supports data preprocessing, exploratory analysis, pathway visualization and metabolite-pathway regression. In summary, Lilikoi v2 is a modern, comprehensive package to enable metabolomics analysis in R programming environment.

Install Lilikoi

Please make sure you have installed these packages (‘impute’, ‘limma’, ‘M3C’, ‘pathifier’, ‘pathview’, ‘preprocessCore’, ‘RCy3’) before installing lilikoi.

# github install for most updated version of lilikoi.
library(remotes)
install_github("lanagarmire/lilikoi2")

# install
# install.packages("lilikoi")
library(lilikoi)

Data description

Plasma breast cancer dataset

dataset2

Load the data with Loaddata function

The recommended format of dataset is csv. Columns are metabolites and rows are samples. Case and control labels should be saved in the second column.

The Loaddata function will load the dataset. For the convenience of further analysis, we extract the output to be a dataset called "Metadata" and a list of compound names called "dataSet".

dt <- lilikoi.Loaddata(file=system.file("extdata", "plasma_breast_cancer.csv", package = "lilikoi"))
Metadata <- dt$Metadata
dataSet <- dt$dataSet

Check the preprocess functions:

lilikoi.preproc_norm(inputdata=Metadata, method="standard")

lilikoi.preproc_knn(inputdata=Metadata, method="knn")

Transform the metabolite names to the HMDB ids using Lilikoi MetaTOpathway function

Lilikoi allows the user to input any kind of metabolite IDs including metabolites names ('name') along with synonyms, KEGG IDs ('kegg'), HMDB IDs ('hmdb') and PubChem IDs ('pubchem'). if the metabolites have a standard names as ID, Lilikoi will match these names among 100k saved database, if there are not any hits, Lilikoi will perform fuzzy matching to find the closest matching for this metabolite. The below table will explain this matching process in more details.

convertResults=lilikoi.MetaTOpathway('name')
Metabolite_pathway_table = convertResults$table
head(Metabolite_pathway_table)

Transform metabolites into pathway using Pathifier algorithm

A specific pathway dysregulation score (PDS) is inferred to measure the abnormity for each sample in each pathway. For each pathway, the samples are mapped in a high dimensional principal component space and a principal curve is constructed along the samples and smoothed. The PDS score measures the distance from the projected dot along the curve to the centroid of normal samples (origin point of the curve).

PDSmatrix= lilikoi.PDSfun(Metabolite_pathway_table)

Using PDSfun, we generate a new matrix which has pathways as columns instead of metaboltes.

# head(t(PDSmatrix))
dim(t(PDSmatrix))

Select the most signficant pathway related to phenotype.

selected_Pathways_Weka= lilikoi.featuresSelection(PDSmatrix,threshold= 0.54,method="gain")

Metabolites-pathway regression

Before running this function, users need to download Cytoscape from http://www.cytoscape？org/download.php. Keep Cytoscape running, then you can connect to R package RCy3. For more info, please refer to vignettes of RCy3. If run multiple times, there will be notifications from Cytoscape asking you to overwrite the former graph.

If you see error like: Error in commandsPOST("vizmap apply styles=\"default\"", base.url = base.url) : No network views selected., please re-run the code and the error will disappear.

lilikoi.meta_path(PDSmatrix = PDSmatrix, selected_Pathways_Weka = selected_Pathways_Weka, Metabolite_pathway_table = Metabolite_pathway_table, pathway = "Alanine, Aspartate And Glutamate Metabolism")

Machine learning

Datasets below are available on the GitHub Site: https://github.com/lanagarmire/lilikoi2/tree/master/inst/extdata.

We updated the lilikoi.machine_learning function so that users can choose which methods they would like to use. The new function of lilikoi.machine_learning is available at: https://github.com/lanagarmire/lilikoi2/blob/master/R/lilikoi.machine_learning.r

dt = lilikoi.Loaddata(file="er_162.csv")
Metadata <- dt$Metadata
dataSet <- dt$dataSet
# library(h2o)
lilikoi.machine_learning(MLmatrix = Metadata, measurementLabels = Metadata$Label,
                              significantPathways = 0,
                              trainportion = 0.8, cvnum = 10, dlround=50,nrun=10, Rpart=TRUE,
                              LDA=TRUE,SVM=TRUE,RF=TRUE,GBM=TRUE,PAM=FALSE,LOG=TRUE,DL=TRUE)

Prognosis

dt <- lilikoi.Loaddata(file = "/Users/rrrrrita/Desktop/lilikoi.temp/JCI71180sd2_processed_lilikoi.csv")

# dt <- lilikoi.Loaddata(file = "JCI71180sd2_processed_tumor.xlsx")

Metadata <- dt$Metadata
dataSet <- dt$dataSet

convertResults=lilikoi.MetaTOpathway('name', chebi=TRUE)
Metabolite_pathway_table = convertResults$table

## Pathway prognosis results - CoxPH
PDSmatrix= lilikoi.PDSfun(Metabolite_pathway_table)

# Extract event and survival time information from the dataset
library(readxl)
jc = read_excel("/Users/rrrrrita/Desktop/lilikoi.temp/JCI71180sd2_processed_tumor.xlsx")

jc <- jc[-27,]
jctime<-jc$TIME
jcevent<-as.numeric(as.factor(jc$EVENT)) - 1


# Pathway dysregulation score information
exprdata = t(as.matrix(PDSmatrix))
exprdata_tumor = exprdata[66:132, ]
exprdata_tumor = exprdata_tumor[-27, ] # The subtracted subject has survival time to be 0, so we deleted it.

# Set up prognosis function arguments
event = jcevent
time = jctime
percent = NULL
exprdata = exprdata_tumor
alpha = 0
nfold = 5
method = "quantile"
cvlambda = "lambda.1se"

# library(survival)
# library(glmnet)
# library(survminer)
lilikoi.prognosis(event, time, exprdata, percent=percent, alpha=0, nfold=5, method="quantile",
          cvlambda=cvlambda,python.path=NULL,coxnnet=FALSE,coxnnet_method="gradient")

## Pathway prognosis results - CoxNNET

# Before running Cox-nnet, users need to provide the directory for python3 and the inst file in lilikoi
path = path.package('lilikoi', quiet = FALSE) # path = "lilikoi/inst/", use R to run
# path = file.path(path, 'inst')

python.path = "/Library/Frameworks/Python.framework/Versions/3.8/bin/python3"


event = jcevent
time = jctime
percent = NULL
exprdata = exprdata_tumor
alpha = 0
nfold = 5
method = "quantile"
cvlambda = NULL
coxnnet = TRUE
coxnnet_method = "gradient"

library(reticulate)

lilikoi.prognosis(event, time, exprdata, percent=percent, alpha=0, nfold=5, method="quantile",
          cvlambda=cvlambda,python.path=python.path,path=path,coxnnet=TRUE,coxnnet_method="gradient")

KEGG (pathview) plot

dt = lilikoi.Loaddata(file=system.file("extdata","plasma_breast_cancer.csv", package = "lilikoi"))
Metadata <- dt$Metadata
dataSet <- dt$dataSet
convertResults=lilikoi.MetaTOpathway('name')
Metabolite_pathway_table = convertResults$table

data_dir=system.file("extdata", "plasma_breast_cancer.csv", package = "lilikoi")
plasma_data <- read.csv(data_dir, check.names=F, row.names=1, stringsAsFactors = FALSE)
sampleinfo <- plasma_data$Label
names(sampleinfo) <- row.names(plasma_data)

metamat <- t(t(plasma_data[-1]))
metamat <- log2(metamat)
grouporder <- c('Normal', 'Cancer')

# install.packages("pathview")
library(pathview)
options(bitmapType='cairo') # Run this line if in linux
lilikoi.KEGGplot(metamat = metamat, sampleinfo = sampleinfo, grouporder = grouporder,
                 pathid = '00250', specie = 'hsa',
                 filesuffix = 'GSE16873', 
                 Metabolite_pathway_table = Metabolite_pathway_table)

Thus, from the KEGGplot function, a graph called "hsa00250.GSE16873.png" has been saved at your working direction.

Duuudude/lilikoi2 documentation built on July 22, 2021, 10:59 a.m.