I am plotting different umaps. I have a part of the code that worked yesterday, however today I get the error message: Error in do.call(c, lapply(2:ncol(nn_idx), function(i) as.vector(rbind(nn_idx[, : 'what' must be a function or character string"
My code is the following:
read.csv (Tabula, row.names=1)%>% as.matrix()-> counts0
t(t(counts0)/colSums(counts0))-> fracs
log(fracs+ 10^-5) -> lfracs
irlba::prcomp_irlba(t(lfracs), n=30)-> pca
colnames (lfracs) -> rownames (pca$x)
rownames (lfracs) -> rownames (pca$rotation)
pca$x %>% head()
set.seed(12345678)
uwot::umap(pca$x, ret_nn=TRUE, spread=5)-> ump
rownames(pca$x)-> rownames(ump$embedding)
as_tibble(ump$embedding)%>%
ggplot + geom_point (aes(x=V1, y=V2), size= .5)+ coord_equal()
Now here is the part where I get the error message:
library(igraph)
louvain_clustering <- function (nn_idx){
do.call (c,lapply (2:ncol(nn_idx), function (i)
as.vector (rbind (nn_idx[,1], nn_idx[,i]))))->edge_list
igraph::make_undirected_graph(edge_list)-> graph
igraph::cluster_louvain(graph)-> clustering
factor(igraph::membership(clustering))
}
louvain_clustering(ump$nn$euclidean$idx)->louvain
as_tibble(ump$embedding)%>%
mutate(cluster=louvain)-> Maus
ggplot(Maus, aes(x=V1, y=V2, col=cluster))+
geom_point(size=.3)+
geom_text(aes(label=cluster), col="black",
data= (Maus %>% group_by (cluster)%>% summarise_all(mean)))+
coord_equal()
The error message appears on the line
louvain_clustering(ump$nn$euclidean$idx)->louvain
and states: Error in do.call(c, lapply(2:ncol(nn_idx), function(i) as.vector(rbind(nn_idx[, : 'what' must be a function or character string"
I really cannot figure out where my mistake lays, hopefully someone can help. Thank you
Maybe you have overwritten the primitive function c ? R lets you do that, I was able to replicate your error bellow, and to fix it you can just remove c and it will revert back to the primitive function, so you can try that, please let me know if it fixes your problem.