mmoisse / pathview Goto Github PK

Currently getting an error when I try to run expand.node=TRUE on a simple single-sample gene expression dataset, when the data works fine with expand.node=FALSE.

Running from a differential expression results table, where the row names are Entrez IDs. Using the "stat" column which is like a t-statistic but negative for negative log-fold-change.

results <- read.csv("results.csv",header=TRUE,row.names=1,check.names=FALSE)
for_pathview <- results$stat
names(for_pathview) <- rownames(results)
pv.out_no_expand = pathview(gene.data = for_pathview,
    pathway.id="04350",
    species = "hsa",
    out.suffix="test",
    kegg.native=FALSE,
    limit=list(gene=10,cpd=10),
    node.sum="max.abs")
pv.out_no_expand = pathview(gene.data = for_pathview,
    pathway.id="04350",
    species = "hsa",
    out.suffix="test",
    kegg.native=FALSE,
    limit=list(gene=10,cpd=10),
    expand.node=TRUE)

With expand.node=TRUE, I get the following error:

Error in combineKEGGnodes(node.data$component[[gn]], gR1, gn) :
Nodes 24, 25 are not part of the graph!

Any help debugging would be much appreciated. Also let me know what if any data you need.

For reference, I wanted to run with expand.node=TRUE because there were a few labels with very different genes within the same label. For example, label ACVR1 had gene ACVR1 itself highly upregulated, but gene BMPR1B (associated gene) highly downregulated. So any of the usual node.sum approaches didn't really seem to fit.