Decide on figures for visualization in Tab 2

[veitveit]

• Think about how to determine "background distributions" -> use decoy hits -> show distributions for decoy hits only
• Cut-off for e.g. FDR score? Take advantage of psm_utils function to calculate FDR (calculate_qvalues) -> slider for filtering

[veitveit]

TODOs for better visualizations

Score and decoy filters: will be become part of the direct column filtering #49

• [ ] Realistic number in maximum limits for filters (might become obsolete with direct columns filtering)
• [ ] Update all figures according to the filters. The first two do not seem to get updated.
• [ ] Coloring of fragment ions types: One could distinguish N-term from C-Term and from internal by each using shades of a different color. d
• [ ] Merge "Distribution of ion types" with "Ion type proportions", e.g. by adding the frequencies to the percentages
• [ ] Change the y label "probability" to more precise label. For example "Number of fragments / all fragments" in "Distribution of m/z values across ion types"
• [ ] In many figures, we show the "proportion" (called "probability"). What about buttons that let the user change between "frequency" and "proportion"?
• [ ] I would prefer to see the intensity in actual numbers, then putting a log-scale on the axis. Also a button to change between relative and a absolute intensity values might be interesting.
• [ ] The x-axis numbers in Fig. 7 are off. Fig. 6-7 should provide actual percentages or the label changed to intensity. This might not need necessary to change a both figures are subject to change.
• [ ] Figure 6 could be substituted by a 2D histogram with terminal and internal ions as axes and the frequency/proportion as colored value. The "other" (better call them "not annotated") are anyways the rest and thus redundant. This provides hopefully a more comprehensive view. Maybe also a figure similar to Arthur's heatmap like view where the terminal ions are at the laterals?
• [ ] Completely revise Fig. 7. I assume it shows the histogram of summed intensities of ion types per spectrum
• [ ] Logo view: Maybe restrict to fixed sequence length? Or show only last/first 3 amino acids? Filtering for the top spectra with most internal ions probably does not bring much as the number of internal ions can also be very high in wrong annotations.