digitalcytometry / ecotyper

EcoTyper is a machine learning framework for large-scale identification of cell states and cellular ecosystems from gene expression data.
Other
181 stars 42 forks source link

Adapt cophenetic coefficient (CC) threshold per cell type #37

Closed LeonHafner closed 2 years ago

LeonHafner commented 2 years ago

Hi there, after processing my scRNA-Seq dataset with the discovery pipeline I checked the plots, especially the rank plot and the cut off for the CC threshold. After visual inspection it seems reasonable to choose different CC thresholds for some cell types than for others as the graph declines differently for each cell type. Is this possible with EcoTyper or is it only possible to set one global threshold for all cell types?

Thanks!

BALuca commented 2 years ago

Hi,

Thanks for your interest in EcoTyper. The current version of EcoTyper only allows to set a global Cophenetic cutoff, shared by all cell types. However, if the automatic selection heuristic is not appropriate for some cell types, one could manually edit file "EcoTyper//Cell_type_specific_genes/Analysis/rank_selection/rank_data.txt" to adjust the number of states per cell type. Assuming that you are following Tutorial 5, after adjusting the number of cell states, you can restart the EcoTyper pipeline from step 4 onwards. This means that you would need to also edit the configuration file field: "Pipeline steps to skip : [1,2,3]". I hope it helps.

Best, The EcoTyper team