Skip to content

Migrate Time-series Analysis from old PROTzilla version #347

@hendraet

Description

@hendraet

Issue description

The old PROTzilla version still has an open PR that tried to implement time-series analysis for PROTzilla: cschlaffner/PROTzilla2#526
This PR includes tooling for loading TMT data, displaying time-series data, and applying different regression methods to it. However, there seemed to be some irregularities in how time metadata was handled in plots, which ultimately prevented this PR from merging. It is advised that the original PR is treated as an outline of features, but code is not blindly copied into this repo.
Please think about how to handle different ways in which time can be stored in metadata (e.g., as date, time point, time elapsed, etc.) to ensure the software stays extensible.

As test data, the proteinGroups.txt from this PRIDE repo can be used (as part of this zip file). A metacsv generated based on the supplementary material can be found here.
There's also a bachelor's thesis including images of all plots, which I can provide upon request.

Acceptance Criteria

  • TMT data can be imported into PROTzilla
  • Time-series data can be displayed as scatter plots where time is displayed on the x-axis
  • Linear regression, RANSAC, and ARIMA can be applied to the data, and the resulting regression outcome can be visualized over the scatter plots.
  • Performance metrics are calculated and displayed for these regression types.
  • Appropriate tests were written

Metadata

Metadata

Assignees

No one assigned

    Labels

    hackathonViable issue for the April 2026 PROTzilla hackathonnew feature

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions