PlateViewer

Quality control visualization tools for arrayed high-content screening experiments (384-well and 96-well plates).

PlateViewer helps microscopists and screening scientists quickly assess image quality across entire plates. It flags out-of-focus wells, reveals intensity patterns such as edge effects or dispensing artifacts, and lets you visually inspect any well or field — all from a browser-based interface.

Requirements

• Python 3.11+
• Images must be TIFF files (.tif) following the naming convention described in Expected File Naming

Installation

Steps:

1. Clone this repository to your local computer:
   • Navigate to the folder where you want to clone (download) this repository.
   • Clone the repo:

git clone https://github.com/aecon/PlateViewer.git

2. Install the package:

cd PlateViewer
conda create -n PlateViewer python=3.11
conda activate PlateViewer
pip install .

On Debian/Ubuntu, the Browse button requires python3-tk, which is not available through pip. In this case install python3-tk as below, and then follow step 2 to re-install PlateViewer.

sudo apt install python3-tk

for development:

For development (editable install — changes take effect immediately):

pip install -e .

reproducing the exact tested environment:

To reproduce the exact tested environment, use the pinned requirements.txt instead:

pip install -r requirements.txt

Usage

Dash (default)

conda activate PlateViewer
plateviewer

The browser opens automatically. Use --port 8051 to change the port.

Streamlit (optional alternative UI)

conda activate PlateViewer  
pip install streamlit
streamlit run plateviewer/app_streamlit.py

Both interfaces expose identical functionality. Dash is the default; Streamlit is an optional alternative that requires no extra configuration beyond pip install streamlit.

Workflow:

1. Enter the plate folder path (or click Browse, or drag-and-drop) and click Load
2. Select a channel and plate format (384-well or 96-well)
3. Optionally enter control well specifications
4. Use the tabs:
   • Random Montage: 4x8 grid of randomly sampled images with well/field labels
   • Control Montage: 4x8 grid sampled from user-specified control wells
   • Single Well: all fields for a chosen well in an auto-sized grid (full images, 4x downsampled)
   • Plate Thumbnails: plate-layout overview with one thumbnail per well (auto-detected center field, hover for well ID)
   • Intensity Heatmap: mean pixel intensity per well (plate layout)
   • Focus Heatmap: two heatmaps — Variance of Laplacian (VoL) and Power Log-Log Slope (PLLS) — with an interpretation guide explaining how to read them together.
5. Click Save All Plots to export all plots as PNGs. By default they are saved to ~/PlateViewer_output/<plate_name>/output/; use the output folder field to choose a different location.

Note: Heatmap results are cached as .npy files under ~/PlateViewer_output/<plate_name>/cache/. Delete them to force recomputation.

Expected File Naming: Images must follow the naming convention below!

PREFIX_ROW - COL(fld FIELD wv WAVELENGTH - CHANNEL).tif

Where PREFIX_ can also be empty. Examples:
*A - 01(fld 1 wv 390 - Blue).tif
*Plate1_A - 01(fld 1 wv 390 - Blue).tif

Control well syntax:

Format	Meaning
A-H:5	Rows A through H, column 5
I-P:13	Rows I through P, column 13
col:1-2	All rows, columns 1 and 2
A01	Single well
A01-A09	Well range within a row

Combine with commas: A-H:5, I-P:13

Functionality

Random Montage

Inspect a subset of randomly selected images from the entire plate, corresponding to the chosen channel.

Single Well

Inspect all fields of a specific well. Enter a well ID (e.g. A05) to see every acquired field arranged in a grid, useful for investigating outliers spotted on heatmaps.

Plate Thumbnails

Bird's-eye view of the entire plate. Each cell shows a thumbnail of the center field for that well. Hover over any thumbnail to see the well ID.

Intensity Heatmap

Mean pixel intensity per well. Reveals systematic patterns such as edge effects or dispensing artifacts.

Focus Heatmaps

Two complementary metrics are shown:

• Variance of Laplacian (VoL) — measures edge/texture content. Higher values indicate sharper images, but VoL also increases with cell confluency and noise.
• Power Log-Log Slope (PLLS) (Bray et al., 2012) — summarizes how quickly spectral power falls off with spatial frequency. More negative values indicate blur; values near zero suggest sharp or noise-dominated images.

License

MIT License
Copyright (c) 2026 E. Athena Economides

If you use this software in your research, please cite it as shown in the "Cite this repository" button, and the CITATION.cff file.

PlateViewer was developed during Athena's Postdoc at:
Prof. Adriano Aguzzi Lab
Institute of Neuropathology
University of Zurich & University Hospital Zurich
Schmelzbergstrasse 12
CH-8091 Zurich
Switzerland