I need to benchmark the reproducibility across replicates of some results on my single-cell RNA-seq datasets, and in order to do so I need to cluster the replicates separately and compare the results. For this task, I would like to be able to input the number of clusters to obtain.
I usually use Seurat's algorithm, but it doesn't allow to specify the number of clusters to compute. SC3 allows that, but it's way too slow on big datasets (considering that I have to test for a lot of values for the number of clusters).
I'm left with SINCERA, that seems to perform fairly good, but the field is so vast that I might be missing something. Does someone have any suggestion?
Dr. Baker’s work in biological informatics and biological clustering analysis at Baylor University might be of interest
Modeling of biologically motivated self-learning equivalent-convolutional recurrent-multilayer neural structures (BLM_SL_EC_RMNS) for image fragments clustering and recognition
- March 2018
- DOI:
- 10.1117/12.2285797
- Badges
- Science topic