I have been running into this issue multiple times, with two different datasets.
Before this error is thrown, phenograph runs until 'Build undirected graph from the weighted links...' is printed.
Then it throws a warning:
'in int2i(as.integer(i), n) :
NAs introduced by coercion to integer range'
And after a few more seconds, it throws the error in the title and stops.
I have checked the following:
-condor object integrity with 'check_IDs()': all good
-anyNA(condor$expr$orig): FALSE
-whether there are duplicated rows in the condor$expr$orig: that was the case, now this is fixed. so no duplicated rows.
-whether the problem is at the knn step, with GPT's suggested code:
library(RANN)
test_knn <- nn2(as.matrix(condor_pooled1$expr$orig), k = 60)
any(is.na(test_knn$nn.idx))
[1]FALSE
I have checked the integrity of the input .fcs files, and that channel names in the expr$orig matrix are the expected ones.
It all looks fine to me.
Could you pinpoint what might be going wrong?
I have been running into this issue multiple times, with two different datasets.
Before this error is thrown, phenograph runs until 'Build undirected graph from the weighted links...' is printed.
Then it throws a warning:
'in int2i(as.integer(i), n) :
NAs introduced by coercion to integer range'
And after a few more seconds, it throws the error in the title and stops.
I have checked the following:
-condor object integrity with 'check_IDs()': all good
-anyNA(condor$expr$orig): FALSE
-whether there are duplicated rows in the condor$expr$orig: that was the case, now this is fixed. so no duplicated rows.
-whether the problem is at the knn step, with GPT's suggested code:
I have checked the integrity of the input .fcs files, and that channel names in the expr$orig matrix are the expected ones.
It all looks fine to me.
Could you pinpoint what might be going wrong?