Hi All
Anyone with experience in the use of DiffCorr package within R (or some alternative) to perform differential network correlation analysis in gene expression ?. I would need some advice and technical help to analyze two sets of gene expression data obtained in two conditions. I cannot use WCGNA because my number of samples is relatively small.
Some help will be appreciated. Thanks in advance. Best
Paco
Dear Francisco
Sorry that do not answer directly on Your question, but just thought that perhaps You can compute simple correlation matrices using "cor" function in R, and then compare these two matrices in R (or even in Excel), using this concept (http://www.philippsinger.info/?p=347)? Sometimes it can be easier to compare two matrices using their representation in a heat-map view (for this is nice to use ggplot2 package in R).
Dear All
Regarding my posted question, I contacted with Atsushi Fukushima, the author of the paper and designer of the package, and I got something that can be interesting for the community. This is a file with an example pipeline for processing data with DiffCorr.
http://sourceforge.jp/projects/sfnet_diffcorr/downloads/AdditionalFile3.txt
I missed this file in the whole DiffCorr site. Hope it can help. Thanks Albert for the suggestion.
Best
Paco
you can also ask https://www.researchgate.net/profile/Ertan_Korkmaz directly
Francisco J. Enguita hi, I need to use the DiffCorr package within R to perform differential network correlation analysis in gene expression (RNA-seq data). Do you know the pipeline to prepare data? What is the input data for DiffCorr? the raw count data? or FPKM?
Bingjie li .... you can use either the raw counts or the normalized data. I would recommend to use the normalized counts to get better results. Good luck
Francisco J. Enguita
Dear Francisco,
Thanks for your reply! I am not sure about the specific procedure as it is my first time to use DiffCorr. The following is my plan. Could you please give me some advice? Can I do the analysis in this way?
1. quantile normalization to the raw counts using DESeq2
2. Filter and remove the low normalized counts
3. log2 transformation
4. as an input of DiffCorr and analysis data follow the script (AdditionalFile3.txt)
It sounds good to me... however I would not use the log2 transformation, since you already had previously quantile normalized data..
This might be an answer too late but if anyone is wondering in future, there's a book chapter here that details the processes involved in using the package.
Chapter Using the DiffCorr Package to Analyze and Visualize Differen...
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Control Theory
- Control Systems