Repository navigation
Waterfall plot for Entity Occurrences #130
Description
Description of feature
An idea for downstream visualisation of comparison between conditions:
Waterfall plot of unique peptides between 2 conditions, with overlapping peptides (non unique). Example figure is attached, in this case y axis = sample prevalence. As we predict the peptides, we can compare the number of entities the peptide is contained in (e.g. bins, contigs, taxids)
Activity
tillenglert commentedon Sep 11, 2024
Bin the entity occurrence to plot less data points (one bin == one entity occurrence)
tillenglert commentedon Dec 10, 2024
Waterfall plot makes sense if it shows high frequency peptides. Frequency in this case would mean a predicted occurrence within one bin/taxa. For these input types the plot generation should be relatively straight forward? (maybe we need to be more stringent in filtering or like I posted above by binning the frequency, to reduce the number of bars -> resources) For assemblies on the other hand it will not be that straight forward, as we do not have information about the taxa/binning in taxa. Still the plot could be generated by using a frequency within contigs, which can be used to compare between communities, but is also dependent on read depth etc and may be biased. Thus either plots are only generated for bins/taxa or the user is warned, that the plot may not show the result that is expected.
Still for large datasets, e.g. full test the plot will generate 2 million bars. Therefore, it may be more helpful to reduce the plot to a table including the result, this could also be a table which filters the highest unique predicted binders (binding score) for each condition.