I would appreciate if someone with extensive knowledge of Model-based clustering specifically on gene expression data, can give a fair comparison on classification potential, Dos and Don't, between available algorithm such as

1.parsimonious Gaussian mixture model (EPGMM) family ,

2. nonparametric model-based clustering, and

3. Gaussian mixture model

All three approaches are implemented in R package; pgmm, pdfCluster, and Mclus, respectively.

Ref.

1 Article Model-Based Clustering of Microarray Expression Data via Lat...

2. Article A novel approach to the clustering of microarray data via no...

3. https://www.stat.washington.edu/mclust/