Plot expression of genes over time

Question

[statfa]

Hi,

What is the name of this plot and how can I draw it?

This plot shows the expression of genes over time.

I have searched on google, but I haven't found its name.

Thanks

[Benn]

These plots can be made in R with matplot()

https://stat.ethz.ch/R-manual/R-devel/library/graphics/html/matplot.html

Answer 1

[Jean-Karim Heriche]

Those are standard line plots/charts which you can do in base R with plot(values, type = 'l') or any number of other plotting functions such as ggplot() or the already mentioned matplot().

[Benn]

Jean-Karim, can you please explain how to use plot(values, type = 'l') on a matrix?

Because I don't get these profile plots then which I do get when I use matplot.

[Jean-Karim Heriche]

In this case, values have to be one of the columns or rows of the matrix. To plot all on the same graph, initialize the plot with plot( ... ,type='n') to not plot anything then loop through the columns to add them with line(). matplot() is essentially a wrapper for this and definitely a shorter way to do this if you want to plot all columns of your matrix on the same graph.

[Benn]

Thanks for the explanation, so the loop method with plot(, type="n") and line() would not be the easiest way (especially not since you want to give each a line a different color or pattern).

[Jean-Karim Heriche]

matplot is just more compact but you still have to give it the parameters needed for plot and line.

[statfa]

Thank you. Can these tools handle replicates in each time point? Or should I obtain the median of normalized expression for replicates in each gene at each time point and then plot them?

[Jean-Karim Heriche]

These are just plotting functions. They will blindly plot what the values you give them. Whether it makes sense to plot all replicates or median values depends on what you want to see.

[statfa]

Oh I see. I was looking for a package that will take the table of read counts with all of its replicates and then draw the plots like the images I showed above.

[Benn]

I am afraid bioinformatics is more than just using one magic package that produces directly publishable images...

I don't know from which study the images are, but I can guess that they first did thorough analysis before they ended up with these figures. It looks like they used clustering first, (like fuzzy clustering or k-means), and then they plotted some of the clusters (using z-scores). I think they have used matplot(), since it looks very similar to plots I have made with matplot().

If the images you posted are from a publication, the methods section should give you information on how to do this.