[Bug] Samples with '+' in ID (e.g., PanCK+) are excluded from Plots tab

[dippindots]

Describe the bug
In the Plots (Beta) tab, samples containing a + character in their Sample ID (e.g., LSP15340_Scan1-003-PanCK+) are being excluded from the visualization when intersecting data profiles. While the Clinical Data tab correctly identifies 96 samples for the selected filters, the Plots tab only picks up 67 samples, seemingly dropping those with the special character.

To Reproduce
Steps to reproduce the behavior:

1. Go to the Ovarian Cancer (Gray Foundation, 2024) study page.
2. Filter ROI Types to Epithelial and Lesion Types to STIC.C (Observe: 96 samples are shown in the clinical table).
3. Click on the Plots (Beta!) tab.
4. Set Horizontal Axis to: CyCIF - Cell Type Fractions -> PanCK+.
5. Set Vertical Axis to: mRNA-Seq Expression GeoMx -> Gene MEX3A.
6. See error: The message "Showing 67 samples with data in both profiles" appears, and samples like LSP15340_Scan1-003-PanCK+ are missing from the plot.

Expected behavior
The plot should include all 96 samples that have data in both profiles. The presence of a + in the sample ID or fraction name should not cause the sample to be filtered out during the data merge.

Screenshots

Additional context
We suspect this is a string-handling or URL-encoding issue where the + sign is not being properly escaped, preventing the frontend from matching the sample IDs across the different data types (CyCIF vs mRNA).

[HarshithaSivalingala]

Hi @dippingdots Can I work on this issue ?

[dippindots]

@HarshithaSivalingala Of course! If you’re interested, please feel free to try reproducing the issue and open a PR with a fix.

[YashhCanCode]

Hello @dippindots
I tried to reproduce this locally and noticed that the UI does not currently allow applying an AND filter for ROI Type = Epithelial and Lesion Type = STIC.C.

The Clinical Data table only supports single-term search and Groups do not restrict samples, so I couldn’t reach the exact “96 samples” state described in the reproduction steps via the UI alone.

However, the underlying issue with sample intersection logic in the Plots (Beta) tab still seems valid, especially for sample IDs containing a + character.

Could you please confirm the intended way to restrict samples for this reproduction.

[HarshithaSivalingala]

Hello @dippindots
cBioPortal/cbioportal-frontend#5408, I have created a PR for this issue. Open to feedback .

[HarshithaSivalingala]

Hi @dippindots dots Any issue with the PR ?