Some feedback and suggestions for Methykit package

[SEvisual]

Hello, developer!

I have been using the Methykit package for analyzing differential methylation, and I want to express my gratitude for developing such a helpful tool for methylation research. However, during my usage, I encountered a few issues and would like to provide some feedback and suggestions:

1.For the functions getMethylationStats() and getCoverageStats(), it would be great if we could visualize the methylation distribution of multiple samples together, such as density plots or boxplots.

2.Regarding the unite() function, it would be useful to have the ability to set the parameter min.per.group as a numeric value or a percentage for each group separately.

3.When using getCorrelation() and clusterSamples(), I noticed that there are NA values present when setting min.per.group in unite() . It would be beneficial to have an option to ignore rows with NA values, as otherwise, it requires rerunning unite().

4.For functions like clusterSamples() and PCASamples(), it would be helpful to change color for group(e.g., red for 1 and blue for 0, owing to 0 represents control samples) and add legend in the plot.

5.In PCA plots, it would be informative to add the title "Proportion of Variance" for the x and y axes labels.

6.Some plots, such as this generated by getCorrelation(), may encounter errors due to large margins (Error in plot.new() : figure margins too large) when there are too many samples. It would be helpful to have an option to directly save such plots by setting parameters accordingly.

7.I checked the Reference Manual and found that the calculateDiffMethDSS() function can be used for differential methylation analysis with the DSS package. Could you please provide a function to convert an object from the Methykit package to an object from the DSS package (methylBase object ---> BSobj object)?

Thank you for considering my feedback. I appreciate your efforts in continuously improving the Methykit package.

Thanks in advance!

[Hannah1746]

I am in huge support for getting "Proportion of Variance" for the x and y axes labels for the PCA plot as well as being able to choose colors and expand the x-axis. Every time I plot a PCA with this software it cuts off the labels so I can't see what is falling where.