Planned Features

[atc3]

Will update this as things are changed

• test separate variance modeling in a structured way
• fix experiment alignment figures image path - needs to be relative to the HTML.
• iteratively run the filters, and check if there are experiments where all observations are filtered out, and then remove them and run the filters again
• fix "sort" issue with pd.concat

q:\anaconda\lib\site-packages\pandas\core\frame.py:6201: FutureWarning: Sorting because non-concatenation axis is not aligned. A future version
of pandas will change to not sort by default.

To accept the future behavior, pass 'sort=True'.

To retain the current behavior and silence the warning, pass sort=False

  sort=sort)

• fix RI columns being wiped when concatenating
• quick and dirty pairwise correlation b/n experiments - to see outliers and warn the user that they should be filtered out
• check the max(PEP) of each raw file, and warn the user if they input a raw file with PEPs that are too low (nothing to boost)
• retention length filtering - raw file specific
• rename output columns
• migrate to config file instead of command-line options
• improve input file-type converting
  • file-type determines column names
  • move filtering blocks into separate functions. file-type determines which functions are run
• pip installable
• violin plot of residual density by RT (RT on x-axis)
• pairwise correlation of RTs - heatmap
• diagnostic figures for the update portion
  • PEP vs PEP.new scatterplot
  • fold change increase in IDs as function of PEP threshold
• validation figures
  • multiple peptides of the same protein - should have the same intensity (measure the CV)
• generate HTML file to view figures
• add and start throwing exceptions
• create entire output directory including all subfolders
• parameter for defining column headers - additional option instead of specifying the file type
• fix experiment exclusion
• optional save alignment parameters
• split up outputs in same way the inputs are split up
  • then remove input_id column
• remove id column as well?
• verbose levels and actually enforce them in code
• additional parameters to select which columns to have
  • default should just be pep_new. maybe have a "diagnostic" flag that includes the other columns?
• logging -> logger
• default retention length filter - (max_retention_time) / 60
• optimize experiment updating
• filter_decoys/contaminants -> include_decoys/contaminants
• add PEP_updated column

FUTURE VERSION

• move off of STAN
• optimize data selection by RT bin, experiment, and peptide. remove as much as possible but retain the same amount of coverage.