search

ShobiStassen / PARC

MIT License
41 stars 11 forks source link

Hyperclustering with PARC #22

Open peterpath opened 1 year ago

peterpath commented 1 year ago

Thank you for this excellent tool!

We are attempting to generate many more clusters than achieved with default parameters. The application is AML-MRD detection by flow cytometry. So far very promising!

What would be your recommended first approach to generate more clusters? 1) increase resolution parameter, 2) decrease jac_std_global, 3) decrease dist_std_local.

Thanks!

ShobiStassen commented 11 months ago

Hi @peterpath ! sorry for not seeing this sooner, somehow I didnt get an email from github. This is too late for you now, but I hope you found what you were looking for. Please feel free to email me directly next time if there is so much delay! You are absolutely right about the parameters, in addition to those you list you can also 1) lower too_big_factor, this is set as 0.3 I believe. a value of 0.1 would for instance force clusters more than 10% of the population to be reclustered further 2) generally lowering the K (knn) value also increases clusters because there is less connectivity in the graph, but doing all of these things at the same time could result in over clustering too.