I am working on single cell RNA-seq data for gene co-expression (GCN) network analysis, I read about the WGCNA and I found it is initially used for microarray datasets. I wonder if it is possible to use for single cell data, If it is not, what are the recommended alternatives?
Dear Yousif A. Algabri
My response to your question is as follows if I am not too late to reply. In order to construct GCN, you need a simple dataframe, which is structured with rows and columns representing genes and samples, respectively. All you have to do is prepare your data from the single cell RNA-seq data accordingly. Once you have input files ready, you can follow the tutorials provided by the WGCNA core team. Thanks
Dear Jagajjit Sahu
Thanks alot for your response. My concern was how good or reliable to use WGCNA for single cell data. As you know it was initially create for microarray data.
Dear Yousif A. Algabri
Initially, it was used with microarray data, but it is an independent method that can be applied to any type of data with a similar aim. In the case of gene expression data, WGCNA aims to create a correlation profile based on expression intensities. It has been used, however, for other high-throughput data, such as transcriptome and metabolome. The same mechanism has been used with gene counts, FPKMs, log fold changes, etc. My personal opinion is that using correlation to GCN is sometimes questionable, especially the values that have been considered when drawing up the correlation profile. However, I still believe it is one of the most effective ways of looking at system-wide insights. Another suggestion would be to use mutual information instead of Pearson's correlation. The Pearsons coefficient often leads to overestimated edges in the GCN. Thanks
Yes, you can use RNA-seq for WGCNA analysis. In fact, most of tools which design for microarray data are also available for RNA-seq data, such as edgeR, DESeq2, and WGCNA. Your need to think two times for the following questions. 1. It is better if your samples are more than 18. 2. it is better if your data were collected from different treatment or tissues. For details, I suggest you read my paper " Genome-wide identification and expression profiling of Alba gene family members in response to abiotic stress in tomato (Solanum lycopersicum L.)".
- Badges
- Science topic