Thank you for this tool!
I am a new user to CONICSmat and just had a few questions about applying it to my own dataset. Ultimately, I am looking to apply CONICSmat to my dataset in order to infer CNVs and separate malignant and non-malignant cells. I have been following the tutorial for CONICSmat on the Oligodendroglioma scRNA-seq dataset.
1) I see that one of the optional inputs for CONICSmat is a table with predicted cell types for each cell. I did not see a step in the tutorial that explained how to incorporate this information when running CONICSmat. I want to run CONICSmat with a matrix of normalized expression counts and a table of predicted cell types, how should I go about doing this?
2) Following the tutorial, I saw that after filtering uninformative noisy regions based on the results of the likelihood ratio test and the BIC for each region, a heatmap of posterior probabilities was generated.
Hi @soerenmueller and @diazlab,
Thank you for this tool! I am a new user to CONICSmat and just had a few questions about applying it to my own dataset. Ultimately, I am looking to apply CONICSmat to my dataset in order to infer CNVs and separate malignant and non-malignant cells. I have been following the tutorial for CONICSmat on the Oligodendroglioma scRNA-seq dataset.
1) I see that one of the optional inputs for CONICSmat is a table with predicted cell types for each cell. I did not see a step in the tutorial that explained how to incorporate this information when running CONICSmat. I want to run CONICSmat with a matrix of normalized expression counts and a table of predicted cell types, how should I go about doing this?
2) Following the tutorial, I saw that after filtering uninformative noisy regions based on the results of the likelihood ratio test and the BIC for each region, a heatmap of posterior probabilities was generated.
hi=plotHistogram(l[,candRegions],suva_expr,clusters=4,zscoreThreshold=4,patients)
With regards to the above line of code, how is the cluster number decided? Is there a method to optimize or choose an appropriate number of clusters?
Thank you in advance.