Entering edit mode
16 months ago
piotto
▴
20
Hello everyone,
What is the process for generating this image? Does a heatmap package create this? Or would there be a classifier or other algorithm generated by the authors?
It seems they tried to see how much of their new cell types possibly with their gene signatures being populated through previous known molecular subtypes.
The image was created by: https://www.nature.com/articles/s41588-019-0566-9
Thank you @barslmn for your response. So, now I may ask how the first figure before this has been generated?
It seems they got four unknown clusters out of their unsupervised analysis (NMF here) and then they tried to see how they will be enriched. But I do not know how they did it and I can see the rows are genes but what are columns? Patients? What of patients? each patient with their raw gene expression data?
That's a heatmap with a dendrogram. You can find more info about it around here like this one: How to generate PCA plot and heatmaps in R
What does the figure caption say about the image?
So, this is it: the description of part a and b is related to the previous image.
So, they have done unsupervised on their data and tried to do hierarchical clustering on their patients. As I do not have their patients's data and If I had I would not try them as i have my own data. I have three clusters from my own data that are tumor cells but unknown. I want to do the exact hierarchical clustering but I only have one patient. So, what would be my columns here as rows are the genes.