| Type: | Package |
| Title: | Metabolomics Personalized Pathway Analysis Tool |
| Version: | 2.1.1 |
| Description: | A comprehensive analysis tool for metabolomics data. It consists a variety of functional modules, including several new modules: a pre-processing module for normalization and imputation, an exploratory data analysis module for dimension reduction and source of variation analysis, a classification module with the new deep-learning method and other machine-learning methods, a prognosis module with cox-PH and neural-network based Cox-nnet methods, and pathway analysis module to visualize the pathway and interpret metabolite-pathway relationships. References: H. Paul Benton http://www.metabolomics-forum.com/index.php?topic=281.0 Jeff Xia https://github.com/cangfengzhe/Metabo/blob/master/MetaboAnalyst/website/name_match.R Travers Ching, Xun Zhu, Lana X. Garmire (2018) <doi:10.1371/journal.pcbi.1006076>. |
| Depends: | R (≥ 3.5.0) |
| Imports: | car, caret, dplyr, gbm, ggplot2, glmnet, h2o, impute, infotheo, limma, M3C, Metrics, MLmetrics, parallel, princurve, pathview, plyr, preprocessCore, pROC, RCy3, reticulate, reshape, RWeka, scales, stats, stringr, survminer, survival |
| License: | GPL-2 |
| Encoding: | UTF-8 |
| RoxygenNote: | 7.2.1 |
| VignetteBuilder: | knitr |
| Suggests: | knitr, rmarkdown |
| NeedsCompilation: | no |
| Packaged: | 2022-10-05 18:08:07 UTC; rrrrrita |
| Author: | Xinying Fang [aut], Yu Liu [aut], Zhijie Ren [aut], Fadhl Alakwaa [aut], Sijia Huang [aut], Lana Garmire [aut, cre] |
| Maintainer: | Lana Garmire <lana.garmire.group@gmail.com> |
| Repository: | CRAN |
| Date/Publication: | 2022-10-05 19:00:02 UTC |
lilikoi.KEGGplot
Description
Visualizes selected pathways based on their metabolites expression data.
Usage
lilikoi.KEGGplot(
metamat,
sampleinfo,
grouporder,
pathid = "00250",
specie = "hsa",
filesuffix = "GSE16873",
Metabolite_pathway_table = Metabolite_pathway_table
)
Arguments
metamat |
metabolite expression data matrix |
sampleinfo |
is a vector of sample group, with element names as sample IDs. |
grouporder |
grouporder is a vector with 2 elements, the first element is the reference group name, like 'Normal', the second one is the experimental group name like 'Cancer'. |
pathid |
character variable, Pathway ID, usually 5 digits. |
specie |
character, scientific name of the targeted species. |
filesuffix |
output file suffix |
Metabolite_pathway_table |
Metabolites mapping table |
Value
Pathview visualization output
Examples
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$table
# data_dir=system.file("extdata", "plasma_breast_cancer.csv", package = "lilikoi")
# plasma_data <- read.csv(data_dir, check.names=FALSE, row.names=1, stringsAsFactors = FALSE)
# sampleinfo <- plasma_data$Label
# names(sampleinfo) <- row.names(plasma_data)
# metamat <- t(t(plasma_data[-1]))
# metamat <- log2(metamat)
# grouporder <- c('Normal', 'Cancer')
# make sure install pathview package first before running the following code.
# library(pathview)
# data("bods", package = "pathview")
# options(bitmapType='cairo')
#lilikoi.KEGGplot(metamat = metamat, sampleinfo = sampleinfo, grouporder = grouporder,
#pathid = '00250', specie = 'hsa',filesuffix = 'GSE16873',
#Metabolite_pathway_table = Metabolite_pathway_table)
A Loaddata Function
Description
This function allows you to load your metabolomics data.
Usage
lilikoi.Loaddata(filename)
Arguments
filename |
file name. |
Value
A data frame named Metadata.
Examples
lilikoi.Loaddata(file=system.file("extdata", "plasma_breast_cancer.csv", package = "lilikoi"))
A MetaTOpathway Function
Description
This function allows you to convert your metabolites id such as names, kegg ids, pubchem ids. into pathways. Metabolites which have not pathways will be excluded from any downstream analysis make sure that you have three database files which are used for exact and fuzzy matching: cmpd_db.rda, syn_nms_db.rda and Sijia_pathway.rda This function was modified version of the name.match function in the below link: https://github.com/cangfengzhe/Metabo/blob/master/MetaboAnalyst/website/name_match.R
Usage
lilikoi.MetaTOpathway(
q.type,
hmdb = TRUE,
pubchem = TRUE,
chebi = FALSE,
kegg = TRUE,
metlin = FALSE
)
Arguments
q.type |
The type of the metabolites id such as 'name', 'kegg', 'hmdb','pubchem' |
hmdb |
if TRUE, match metabolites id to the HMDB database. |
pubchem |
if TRUE, match metabolites id to the PubChem database. |
chebi |
if TRUE, match metabolites id to the ChEBI database. |
kegg |
if TRUE, match metabolites id to the KEGG database. |
metlin |
if TRUE, match metabolites id to the METLIN database. |
Value
A table showing the convertion results from metabolites ids to ids in different metabolomics databases and pathway ids and names.
Examples
dt <- lilikoi.Loaddata(file=system.file("extdata",
"plasma_breast_cancer.csv", package = "lilikoi"))
Metadata <- dt$Metadata
dataSet <- dt$dataSet
# Metabolite_pathway_table=lilikoi.MetaTOpathway('name')
A PDSfun Function
Description
This function allows you to compute Pathway Desregulation Score deriving make sure that you have the below database for the metabolites and pathway list: meta_path.RData
Usage
lilikoi.PDSfun(qvec)
Arguments
qvec |
This is the Metabolite_pathway_table from MetaTOpathway function. This table includes the metabolites ids and the its corssponding hmdb ids |
Value
A large matrix of the pathway deregulation scores for each pathway in different samples.
References
Nygård, S., Lingjærde, O.C., Caldas, C. et al. PathTracer: High-sensitivity detection of differential pathway activity in tumours. Sci Rep 9, 16332 (2019). https://doi.org/10.1038/s41598-019-52529-3
Examples
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$table
# PDSmatrix= lilikoi.PDSfun(Metabolite_pathway_table)
Exploratory analysis
Description
Performs source of variation test and build PCA and t-SNE plots to visualize important information.
Usage
lilikoi.explr(data, demo.data, pca = FALSE, tsne = FALSE)
Arguments
data |
is a input data frame for analysis with sample ids as row names and metabolite names or pathway names as column names. |
demo.data |
is a demographic data frame with sample ids as row names, sample groups and demographic variable names as column names. |
pca |
if TRUE, PCA plot will be out. |
tsne |
if TRUE, T-SNE plot will be out. |
Value
Source of variation test results and PCA and t-SNE plot
Examples
# lilikoi.explr(data, demo.data, pca=TRUE, tsne=FALSE)
A featuresSelection Function
Description
This function allows you to reduce the pathway diemsion using xxxx
Usage
lilikoi.featuresSelection(PDSmatrix, threshold = 0.5, method = "info")
Arguments
PDSmatrix |
from PDSfun function |
threshold |
to select the top pathways |
method |
information gain ("info") or gain ratio ("gain") |
Value
A list of top metabolites or pathways.
Examples
dt <- lilikoi.Loaddata(file=system.file("extdata",
"plasma_breast_cancer.csv", package = "lilikoi"))
Metadata <- dt$Metadata
dataSet <- dt$dataSet
# Metabolite_pathway_table=lilikoi.MetaTOpathway('name')
# PDSmatrix= lilikoi.PDSfun(Metabolite_pathway_table)
# selected_Pathways_Weka= lilikoi.featuresSelection(PDSmatrix,threshold= 0.50,method="gain")
A machine learning Function
Description
This function for classification using 8 different machine learning algorithms and it plots the ROC curves and the AUC, SEN, and specificty
Usage
lilikoi.machine_learning(
MLmatrix = PDSmatrix,
measurementLabels = Label,
significantPathways = selected_Pathways_Weka,
trainportion = 0.8,
cvnum = 10,
dlround = 50,
nrun = 10,
Rpart = TRUE,
LDA = TRUE,
SVM = TRUE,
RF = TRUE,
GBM = TRUE,
PAM = TRUE,
LOG = TRUE,
DL = TRUE
)
Arguments
MLmatrix |
selected pathway deregulation score or metabolites expression matrix |
measurementLabels |
measurement label for samples |
significantPathways |
selected pathway names |
trainportion |
train percentage of the total sample size |
cvnum |
number of folds |
dlround |
epoch number for the deep learning method |
nrun |
denotes the total number of runs of each method to get their averaged performance metrics |
Rpart |
TRUE if run Rpart method |
LDA |
TRUE if run LDA method |
SVM |
TRUE if run SVM method |
RF |
TRUE if run random forest method |
GBM |
TRUE if run GBM method |
PAM |
TRUE if run PAM method |
LOG |
TRUE if run LOG method |
DL |
TRUE if run deep learning method |
Value
Evaluation results and plots of all 8 machine learning algorithms, along with variable importance plots.
Examples
dt = lilikoi.Loaddata(file=system.file("extdata","plasma_breast_cancer.csv", package = "lilikoi"))
Metadata <- dt$Metadata
# lilikoi.machine_learning(MLmatrix = Metadata, measurementLabels = Metadata$Label,
# significantPathways = 0,
# trainportion = 0.8, cvnum = 10, dlround=50,Rpart=TRUE,
# LDA=FALSE,SVM=FALSE,RF=FALSE,GBM=FALSE,PAM=FALSE,LOG=FALSE,DL=FALSE)
Metabolite-pathway regression
Description
Performs single variate linear regression between selected pathways and each of their metabolites. Output the network plot between pathways and metabolties.
Usage
lilikoi.meta_path(
PDSmatrix,
selected_Pathways_Weka,
Metabolite_pathway_table,
pathway = "Pyruvate Metabolism"
)
Arguments
PDSmatrix |
Pathway deregulation score matrix |
selected_Pathways_Weka |
Selected top pathways from the featureSelection function |
Metabolite_pathway_table |
Metabolites mapping table |
pathway |
interested pathway name |
Value
A bipartite graph of the relationships between pathways and their corresponding metabolites.
An imputation function.
Description
This function is used to preprocess data via knn imputation.
Usage
lilikoi.preproc_knn(inputdata = Metadata, method = c("knn"))
Arguments
inputdata |
An expression data frame with samples in the rows, metabolites in the columns |
method |
The method to be used to process data, including |
Value
A KNN imputed dataset with samples in the rows, metabolites in the columns.
Examples
dt <- lilikoi.Loaddata(file=system.file("extdata",
"plasma_breast_cancer.csv", package = "lilikoi"))
Metadata <- dt$Metadata
dataSet <- dt$dataSet
lilikoi.preproc_knn(inputdata=Metadata, method="knn")
A Normalization function.
Description
This function is used to preprocess data via normalization. It provides three normalization methods: standard normalization, quantile normalization and median fold normalization. The median fold normalization is adapted from http://www.metabolomics-forum.com/index.php?topic=281.0.
Usage
lilikoi.preproc_norm(
inputdata = Metadata,
method = c("standard", "quantile", "median")
)
Arguments
inputdata |
An expression data frame with samples in the rows, metabolites in the columns |
method |
The method to be used to process data, including standard normalization (standard), quantile normalization (quantile) and median fold normalization (median). |
Value
A normalized dataset with samples in the rows, metabolites in the columns.
Examples
dt <- lilikoi.Loaddata(file=system.file("extdata",
"plasma_breast_cancer.csv", package = "lilikoi"))
Metadata <- dt$Metadata
dataSet <- dt$dataSet
lilikoi.preproc_norm(inputdata=Metadata, method="standard")
Pathway-based prognosis model
Description
Fits a Cox proportional hazards regression model or a Cox neural network model to predict survival results.
Usage
lilikoi.prognosis(
event,
time,
exprdata,
percent = NULL,
alpha = 1,
nfold = 5,
method = "median",
cvlambda = "lambda.1se",
python.path = NULL,
path = NULL,
coxnnet = FALSE,
coxnnet_method = "gradient"
)
Arguments
event |
survival event |
time |
survival time |
exprdata |
dataset for penalization, with id in the rownames and pathway or metabolites names in the column names. |
percent |
train-test separation percentage |
alpha |
denote which penalization method to use. |
nfold |
fold number for cross validation |
method |
determine the prognosis index, "quantile", "quantile" or "ratio". |
cvlambda |
determine the lambda for prediction, "lambda.min" or "lambda.1se". |
python.path |
saved path for python3 |
path |
saved path for the L2cross_nopercent.py and L2cross.py files in lilikoi |
coxnnet |
if TRUE, coxnnet will be used. |
coxnnet_method |
the algorithm for gradient descent. Includes standard gradient descent ("gradient"), Nesterov accelerated gradient "nesterov" and momentum gradient descent ("momentum"). |
Value
A list of components:
c_index |
C-index of the Cox-PH model |
difftest |
Test results of the survival curve difference test |
survp |
Kaplan Meier plot |
Examples
# inst.path = path.package('lilikoi', quiet = FALSE) # path = "lilikoi/inst/", use R to run
# inst.path = file.path(inst.path, 'inst')
# python.path = "/Library/Frameworks/Python.framework/Versions/3.8/bin/python3"
# Prepare survival event, survival time and exprdata from your dataset.
# lilikoi.prognosis(event, time, exprdata, percent=NULL, alpha=0, nfold=5, method="median",
# cvlambda=NULL,python.path=NULL, path=inst.path, python.path=python.path,
# coxnnet=FALSE,coxnnet_method="gradient")