dim(X) must have a positive length

[KeitaSaeki]

Hi,

I encountered the above issue.
When I used for the first time, I analyzed 143 cells in comparison with 47 cells, in which 15017 genes are expressed, and worked well. But when I compared the other 724 cells vs 143 cells with the same number of genes, I have got the following error message.

Remove genes that all are zeros...
done
Warning messages:
1: fitted probabilities numerically 0 or 1 occurred
2: fitted probabilities numerically 0 or 1 occurred
3: fitted probabilities numerically 0 or 1 occurred
Error in apply(x2[which(x != 0), ], 2, FUN = function(x, b) { :
dim(X) must have a positive length
Calls: FunImpute -> sapply -> lapply -> FUN -> apply

Is the issue due to the number of cells that I included?
Can we solve this issue?

Thanks for your help in advance.

Best,

Keita

[tianyu-github]

Hi,

I encountered the above issue.
When I used for the first time, I analyzed 143 cells in comparison with 47 cells, in which 15017 genes are expressed, and worked well. But when I compared the other 724 cells vs 143 cells with the same number of genes, I have got the following error message.

Remove genes that all are zeros...
done
Warning messages:
1: fitted probabilities numerically 0 or 1 occurred
2: fitted probabilities numerically 0 or 1 occurred
3: fitted probabilities numerically 0 or 1 occurred
Error in apply(x2[which(x != 0), ], 2, FUN = function(x, b) { :
dim(X) must have a positive length
Calls: FunImpute -> sapply -> lapply -> FUN -> apply

Is the issue due to the number of cells that I included?
Can we solve this issue?

Thanks for your help in advance.

Best,

Keita

It is not because of the number of the cells you used. The error occured in the function of imputation, because all the expressions are zeros when imputing a small group of the zero expressions. I guess the problems is teh data is so sparse and the ratio is too low to choose the "genes_use" if the function of "idfyUsegene(data,databinary,condition,ratio = 0.5)" in line 28 in the example.R. Could you have a test that not to use the FunImpute(), and set the "Hur_gene = NULL" in line 42?

[KeitaSaeki]

It worked! Thank you!

Keita

Hi,
I encountered the above issue.
When I used for the first time, I analyzed 143 cells in comparison with 47 cells, in which 15017 genes are expressed, and worked well. But when I compared the other 724 cells vs 143 cells with the same number of genes, I have got the following error message.
Remove genes that all are zeros...
done
Warning messages:
1: fitted probabilities numerically 0 or 1 occurred
2: fitted probabilities numerically 0 or 1 occurred
3: fitted probabilities numerically 0 or 1 occurred
Error in apply(x2[which(x != 0), ], 2, FUN = function(x, b) { :
dim(X) must have a positive length
Calls: FunImpute -> sapply -> lapply -> FUN -> apply
Is the issue due to the number of cells that I included?
Can we solve this issue?
Thanks for your help in advance.
Best,
Keita

It is not because of the number of the cells you used. The error occured in the function of imputation, because all the expressions are zeros when imputing a small group of the zero expressions. I guess the problems is teh data is so sparse and the ratio is too low to choose the "genes_use" if the function of "idfyUsegene(data,databinary,condition,ratio = 0.5)" in line 28 in the example.R. Could you have a test that not to use the FunImpute(), and set the "Hur_gene = NULL" in line 42?