Hello,
I am trying to use the permutation testing functionality from NetCoMi::netCompare but I am running into some problems with several of my datasets. I think the error is coming from around line 865 of NetCoMi::netCompare based on the final error, the permutation running, and the progress bar not closing before the error is reported but I do not know the helper functions well enough to have a more informed guess. My input matrices are not particularly sparse and do not have any rows/columns that sum to 0 but maybe during the permutation tests things shuffle and make combinations that cause problems. With this toy example the errors all seem to have happened on the 7th column, which is pretty sparse in d1. Thank you for your time and for your work on NetCoMi!
Edit: I have not been able to get around this error using the zeroMethod and zeroPar arguments to netConstruct to pass arguments to the zCompositions options but I may be missing something in those. Using the pseudo/pseudoZO options giving a pseudocount of 1 let my example data run, albeit more slowly. Since that works I'm going to close this issue.
Hi,
Thanks for the detailed explanations. Unfortunately, I don't have the time to take a closer look at this issue this week. Even though it's closed, I'll try to find out what's wrong with zCompositions.
Hello, I am trying to use the permutation testing functionality from
NetCoMi::netCompare
but I am running into some problems with several of my datasets. I think the error is coming from around line 865 ofNetCoMi::netCompare
based on the final error, the permutation running, and the progress bar not closing before the error is reported but I do not know the helper functions well enough to have a more informed guess. My input matrices are not particularly sparse and do not have any rows/columns that sum to 0 but maybe during the permutation tests things shuffle and make combinations that cause problems. With this toy example the errors all seem to have happened on the 7th column, which is pretty sparse ind1
. Thank you for your time and for your work on NetCoMi!Edit: I have not been able to get around this error using the zeroMethod and zeroPar arguments to
netConstruct
to pass arguments to the zCompositions options but I may be missing something in those. Using the pseudo/pseudoZO options giving a pseudocount of 1 let my example data run, albeit more slowly. Since that works I'm going to close this issue.