Hello, I think I may have found a bug in the code for aveLogCPM.default in the edgeR package. Near the end of the function, the variable "prior.count.scale" is conditionally assigned to, and never subsequently used. I assume this is a typo and the variable name is supposed to be "prior.count.scaled" instead. Thus, this code is essentially a no-op. I'm not sure what the implications of skipping this line are or when this code would be executed. The line in question is this: if(is.null(dim(prior.count.scaled))) prior.count.scale <- matrix(1,nrow(y),1) %*% prior.count.scaled I downloaded the current development tarball from the Bioconductor site, and this also seems to have the same issue. On a side, note, the edgeR page lists 4 maintainers. Which one is the best one to email about bug reports, or should I cc all of them? Thanks, -Ryan Thompson
[Bioc-devel] Possible bug in edgeR::aveLogCPM.default?
2 messages · Ryan, Gordon K Smyth
Hi Ryan,
On Fri, 11 Jul 2014, Ryan C. Thompson wrote:
Hello, I think I may have found a bug in the code for aveLogCPM.default in the edgeR package. Near the end of the function, the variable "prior.count.scale" is conditionally assigned to, and never subsequently used. I assume this is a typo and the variable name is supposed to be "prior.count.scaled" instead. Thus, this code is essentially a no-op. I'm not sure what the implications of skipping this line are or when this code would be executed. The line in question is this: if(is.null(dim(prior.count.scaled))) prior.count.scale <- matrix(1,nrow(y),1) %*% prior.count.scaled
Yes, you're right. Thanks for letting us know. Will fix. Typo was introduced last November.
I downloaded the current development tarball from the Bioconductor site, and this also seems to have the same issue. On a side, note, the edgeR page lists 4 maintainers. Which one is the best one to email about bug reports, or should I cc all of them?
The author of each function is commented in the code, and also identified in the Rd file. There is no need to mail minor bug reports to bioc-devel. Best wishes Gordon
Thanks, -Ryan Thompson
______________________________________________________________________
The information in this email is confidential and intend...{{dropped:4}}