Dependency Conflict in GSVA Package of Commpath: matrixStats Version Incompatibility
Opened this issue · 3 comments
Issue Description
I encountered an issue related to the matrixStats package, which is a dependency of the GSVA package.
When using the scorePath(obj, method = "gsva") function in the GSVA package, I encountered an error. This appears to be due to a recent update in the matrixStats package. The latest version of matrixStats no longer supports the argument usenames=NA. This change has led to compatibility issues with the GSVA package, specifically affecting the functionality of the scorePath function.
Temporary Solution
To circumvent this issue, I found that rolling back to version 1.0.0 of the matrixStats package resolves the problem. However, this is not a sustainable solution as it prevents using the latest features and fixes of the matrixStats package.
Suggested Action
It would be beneficial for the Commpath team to address this compatibility issue. This might involve updating the GSVA package to be compatible with the latest version of matrixStats, or providing guidelines on handling this dependency conflict.
Thank you for your attention to this matter. I appreciate the work you do and look forward to a resolution that allows for the continued use of both the GSVA package and the latest version of matrixStats.
I'm glad to hear that CommPath has been helpful for your work! Thank you for your kind words and for sharing your feedback.
I have tested the scorePath function using the latest versions of GSVA (v1.50.0) and matrixStats (v1.2.0) on R (v4.3.1), and the function runs without any issues. To replicate the error, could you please provide an example dataset along with the corresponding code?
Best
Additional Information: I would like to point out that I am currently using R version 4.1. This may be contributing to the issue where Bioconductor does not support the latest version of matrixStats, leading to the dependency conflict.
Due to the limitation of my R version, I am unable to update the GSVA package to the latest version that might avoid this bug. This seems to be a key reason for the issue I am encountering. If possible, I suggest including specific requirements for the R version in the package documentation to assist other users who might face a similar problem.
Here for more information.
I hope this information is helpful in resolving the issue. Thank you once again for your support and assistance.
Best regards,
Thank you for your carful inspection and suggestions!
We're considering to check the issue and update it soon.
Best,