What feature would you like to propose?
I'm made aware of the availability of FlowSOM in Python
We should support this because FlowSOM is ubiquitously used in CyTOF workflow.
Is your feature request related to a current problem? Please describe.
We currently do not have a clustering algorithm on hand. This is much needed.
How do you envision this feature to be implemented?
I've glanced at the documentation, but a problem is that FlowSOM needs either FCS or AnnData, neither of which we support natively. But it may possible to work with csv file. If the latter is the case, we can possibly work around that by temporarily cache the dataset, which is ugly (IO bottleneck, wasted step, etc.) but possible.
Are there any existing alternatives or workarounds?
Not really. Currently, we don't offer an interface to AnnData or FCS. So, users need to use standard clustering algorithms such as those from sklearn.
What feature would you like to propose?
I'm made aware of the availability of FlowSOM in Python
We should support this because FlowSOM is ubiquitously used in CyTOF workflow.
Is your feature request related to a current problem? Please describe.
We currently do not have a clustering algorithm on hand. This is much needed.
How do you envision this feature to be implemented?
I've glanced at the documentation, but a problem is that FlowSOM needs either FCS or AnnData, neither of which we support natively. But it may possible to work with csv file. If the latter is the case, we can possibly work around that by temporarily cache the dataset, which is ugly (IO bottleneck, wasted step, etc.) but possible.
Are there any existing alternatives or workarounds?
Not really. Currently, we don't offer an interface to AnnData or FCS. So, users need to use standard clustering algorithms such as those from sklearn.