You can run cluster analysis using SPSS, Genstat or Darwin 5.0 statistical software using the hierachical cluster analysis option

You can apply co-clustering treatments and samples with together by bi-clustering or blockclustering R packages. 

Dear

Note that edgeR and DESeq require readcounts for downstream analysis. But to use edgeR or DESeq, you can make of  htseq-count to get the raw counts per gene/exon and feed that data into edger/DESeq.

Please check the link: https://insilicodb.com

Regards

https://www.researchgate.net/post/How_can_I_do_clustering_with_RNA-Seq_data

Hi Ada

Could you elaborate what you mean by by identifying linear trends - in that way we might be able to better help you :-)

confused and confusing question. Please elaborate. Clustering means to obtain a partition of your objects into groups putting together objects that are more similar. This can be done in a huge number of ways and methods. In common to all:

first point to understand is the direction on which you want to perform the clustering. Do you want to cluster the genes, or the samples? In case of genes, you have to consider the number of samples that you have, because if too small, it does not really make any sense to do the clustering, especially with some of the distance functions you can use. For instance, with three samples, you can cluster by correlation, but you are in the very minimum number to get a meaningful output. If you want to cluster the samples instead, you should first identify a set of genes showing enough variations (differentially expressed genes can be a good starting point, or alternatively, you can select genes on the basis of their variability in the different samples without performing a significance analysis of the changes). Then you have to normalize the data, which is not really necessary if you calculate correlations, but it is instead fundamental e.g. with euclidean distance. and at end run your clustering algorithm.

You can use for instance the heatmap or heatmap2 functions in R-cran/R-studio, or clustergram in matlab that do not require deep knowledge of the programming language/all parameters behind the tool.

 Dear Matteo,

Thanks a lot for exact and clear answer.