Skip to content

Condition Tables

Condition tables let you attach experimental metadata — such as drug treatments, concentrations, or replicate numbers — to specific wells in your plate. This information is stored in the OME-Zarr output and can be used by downstream analysis tasks.

Format

A condition table is a CSV file with the following structure:

Required Columns

Column Type Description
row string Well row letter (A, B, C, ...)
column or col integer Well column number (1, 2, 3, ...)

Optional Columns

Column Type Description
acquisition integer Acquisition ID. If present, conditions are filtered by acquisition.
any other column string, int, float, or bool Custom experimental metadata.

Column name matching

Column names are matched case-insensitively. Both Row and row work. For the column number, both column and col are accepted.

Example

row,column,acquisition,drug,concentration,replicate
A,1,0,DMSO,0,1
A,2,0,DrugA,0.1,1
A,3,0,DrugA,1.0,1
B,1,0,DMSO,0,2
B,2,0,DrugA,0.1,2
B,3,0,DrugA,1.0,2

You can also have multiple rows per well to represent multiple conditions applied to the same well:

row,column,acquisition,drug,concentration,replicate
C,11,0,drugA,0.2,1
C,11,0,drugB,,2

Data Type Rules

Each custom column must contain values of a single type. The converter auto-detects the type:

  • Strings: Any text values. Empty strings, NA, N/A, and Na are treated as missing (null).
  • Numbers: Integers or floats. NaN values are preserved.
  • Booleans: true / false (case-sensitive).

Mixing types within a single column (e.g., some rows are strings, others are numbers) will cause an error.

Usage

To use a condition table, set the condition_table_path in the advanced options of your acquisition:

Note

The condition table path must be absolute.

What Happens

When a condition table is provided:

  1. For each well in the plate, the converter looks up matching rows in the CSV (by row + column, optionally filtered by acquisition).
  2. The matched metadata columns are attached as attributes to the well's image tiles.
  3. These attributes are written into:
    • At the OME-Zarr image level (each image contains the conditions that apply to it)
    • At the plate level (aggregated table of all image conditions)
    • Returned to the Fractal serve to be stored in datasets attributes