Entering edit mode
8.6 years ago
randalljellis
▴
90
Hello,
STRING_db is a Bioconductor library used to assess protein interactions. It relies on another library called iGraph that is used to cluster proteins with a get_clusters function. When I run it on a list of genes like so, I get an error:
clustersList <- string_db$get_clusters(mydata)
Error in .Call("R_igraph_community_fastgreedy", graph, as.logical(merges), :
At community.c:769 : `steps' to big or `merges' matrix too short, Invalid value
I've read that this is a common error with iGraph and found a workaround online for python, but not for R. I'm not sure what to do. If anyone can provide any insight on this I would greatly appreciate it!
This can be produced if your graph is not fully connected. The workaround is to first extract the connected components then apply clustering to each separately. In my experience, regardless of the clustering algorithm, it's better to work with connected components separately.
Thank you for the insight, much appreciated!