14 months ago by
The Alan Turing Institute
Indeed, the current release of BASiCS (BioC 3.6) cannot be used without spike-ins. Recently, we extended the method to the no-spikes case and that is available in BioC devel (see also https://www.biorxiv.org/content/early/2017/12/21/237214).
However, I am afraid that BASiCS is *not* suitable for clustering analyses. This is because BASiCS has been designed in the context of supervised experimental designs only (e.g. case/control). Applying BASiCS in a clustering context is not advised and therefore I would refer you to other libraries that are specifically designed for clustering (e.g. SC3).
My apologies for not being more helpful this time.