Heat map of DE genes identified in Graph based clustering of ScRNA-Seq

Question

[rob.costa1234]

I have scRNA-seq expression matrix (read counts) after graph based clustering. I want to show the clusteras as a heatmap shown in figure 1E or 1F reported in this paper https://www.cell.com/cell-reports/abstract/S2211-1247(18)30339-5 or figure can be accessed directly @ https://ibb.co/mwf16U I want to show this kind of heat map along with tSNE plot. The heat map tools MeV4, R does not give me specifically this kind of figure. In the figure there are specific number of cells (Corresponding to width of red color ) that express signature genes of one particular cluster.

Alternative suggestion of the tools to show this kind of heat map from graph based clusters are also welcome.

[WouterDeCoster]

See How to add images to a Biostars post
You make people search for your figure before they can help you. Make it as easy as possible.

You also didn't tell us which data you have.

And please select a more descriptive title for your question. "to draw a figure" doesn't tell us what this is about.

[Michael]

(E and F) Heat maps of MC normalized signal show MC subtypes changes by top genes (columns) for individual MC subtype cells (rows) in normal (E) and fibrotic (F) αSMA-GFP;Tbx4-Cre;Rosa26-tdTomato lung samples.

So you just want to make heat-maps? There are tons of tools (including R) and documentation on the internet to draw heat maps. Please search for posts on Biostars or provide a more specific description of your problem. For ease of mind, just click on the heatmap tag to get 8 pages of heatmap questions. Or see Heatmaps in R

[Michael]

The heat map tools MeV4, R does not give me specifically this kind of figure.

By using which code exactly? I am quite sure the figures can be made with R heatmap.2 or ggplot.

In the figure there are specific number of cells (Corresponding to width of red color ) that express signature genes of one particular cluster.

This can be achieved by using parameters ColSideColors, RowSideColors, and turning off row and column re-ordering.

[lakhujanivijay]

As suggested by WouterDeCoster, you should immediately change the title for the question

Answer 1

[Kristin Muench]

Not sure if this is exactly what you're looking for, but I highly recommend the package pheatmap(). I made heatmaps in ggplot2 for years, and they came out fine, but it was a little annoying to use - e.g. I found it tricky to order rows and columns a certain way. Pheatmap() makes pretty heatmaps, plots factors on the side (e.g. if you wanted to plot multiple clusters and visualize what block belongs to which cluster), clusters the rows/columns for you (or not), and so forth.

Also recommend changing the default colors by using RColorBrewer with the color = flag in pheatmap, e.g. color= colorRampPalette(rev(brewer.pal(n = 7, name ="YlOrRd")))(100)