This minimum example from the documentation works for multiple columns:
library(pathview)
sim.cpd.data=sim.mol.data(mol.type="cpd", nmol=3000)
sim.cpd.data2 = matrix(sample(sim.cpd.data, 18000,
replace = T), ncol = 6)
i <- 3
pv.out <- pathview(gene.data = gse16873.d[, 1:6],
pathway.id = "00640",
species = "hsa",
keys.align = "y",
kegg.native = T,
match.data = F,
multi.state = T,
same.layer = T
)
This section of code from NewFullRNASeqPipeline.R works for a single condition:
#__________Python Enrichment_________#
percentiles <- c(.99, .95, .90)
test <- keggEnrichmentOnPercentiles(formatedCSVThereshold0, percentiles, metric="score", outputDir=OutputFileDirectory)
pathways <- py_GetAllPathwaysOfSig(test, p_cutoff=1.0)
pathways <- pathways[pathways != 'cme01100' & pathways != 'cme01110'] # Exclude these larger diagrams
keggDataFrame <- py_getKEGGDataframe(formatedCSVThereshold0)
colnames(keggDataFrame)
# kegg_LFC <- keggDataFrame[,c("LFC_LP1hVRM1h", "LFC_LP2hVRM2h", "LFC_LP24hVRM24h", "LFC_LP49hVRM49h")]
scoreFilteredDataFrame = keggDataFrame['Score_LPvRM' >= 33] # 33 comes from LP manuscript; not for general purpose
kegg_LFC <- scoreFilteredDataFrame[,c(paste("LFC_", treated, "v", control, sep="")), drop=FALSE]
#Flip sign so that legend shows green as positive and red negative.
kegg_LFC <- -1*kegg_LFC
#pathview(gene.data = kegg_LFC, pathway.id = "cme03030", species = "cme", gene.idtype = "KEGG",plot.col.key=FALSE)
#kegg_LFC["CYME_CMK133C",]
pv.out.list <- sapply(
pathways,
function(pid) pathview(
gene.data = kegg_LFC,
pathway.id = pid,
species = "cme",
gene.idtype = "KEGG",
plot.col.key=FALSE
)
)
The following approach uses the LPRM data, but the same args as the minimum working example. But it doesn't work. The pathway diagrams are generated, but assigned values are all white.
source("NewFullRNAseqPipeline_metadata.R")
pathways <- scan("../Annotations/Annotations/kegg_pathways.csv", character())
pathways <- pathways[pathways != 'cme01100' & pathways != 'cme01110'] # Exclude these larger diagrams
timepoints <- unique(coldata$timepoint)
t_dfs <- data.frame()
c <- 0
for (t in timepoints) {
c <- c + 1
print(t)
t_col <- coldata[coldata$timepoint == t,]
t_cts <- cts[grepl(t, colnames(cts))]
dds <-DESeqDataSetFromMatrix( countData = t_cts,
colData = t_col,
design = configDesign)
dds <- DESeq(dds)
res <- results(dds)
t_res <- as.data.frame(res)[c("log2FoldChange", "padj")]
t_res$log2FoldChange <- ifelse(t_res$padj < 0.01, t_res$log2FoldChange, 0)
t_res <- t_res['log2FoldChange']
names(t_res)[names(t_res) == 'log2FoldChange'] <- paste(t, c, sep='_')
if (nrow(t_dfs) == 0){
t_dfs <- t_res
} else {
t_dfs = merge(t_dfs, t_res, by="row.names", all=TRUE)
rownames(t_dfs) <- t_dfs$Row.names
t_dfs <- subset(t_dfs, select=-c(Row.names))
}
}
kegg_LFC <- t_dfs * -1
kegg_LFC <- na.omit(kegg_LFC)
pv.out.list <- sapply(pathways,
function(pid) pathview(
gene.data = kegg_LFC,
pathway.id = pid,
species = "cme",
gene.idtype = "KEGG",
limit = c(min(kegg_LFC), max(kegg_LFC)),
keys.align = "y",
kegg.native = T,
match.data = F,
multi.state = T,
same.layer = T,
plot.col.key=FALSE
)
)
This minimum example from the documentation works for multiple columns:
This section of code from
NewFullRNASeqPipeline.Rworks for a single condition:The following approach uses the LPRM data, but the same args as the minimum working example. But it doesn't work. The pathway diagrams are generated, but assigned values are all white.