Marker Objects¶

Summary¶

Marker objects wrap imported measurement tables. They are the values stored in experiment.data and batch.data, and each one exposes its table as .df.

Most users inspect marker objects rather than constructing them directly. The full import pipeline chooses the object type from the source folder:

Object	Typical Source	Role
`Attribute`	ROI properties, behavior, generic tables, flat CSVs	Generic table wrapper with cleaned columns and condition metadata.
`Antibody`	ROI intensity tables	Marker intensity table with marker-prefixed columns.
`cellMarker`	Cell marker object tables	Object table without colocalisation summary generation.
`objectMarker`	Object marker tables	Object table with colocalisation count and closest-distance support.

How To Create It¶

Usually through Experiment or Batch processing:

from PyFLASH import create_batch

batch = create_batch(
    "example",
    groups,
    "/path/to/output-folder",
    experiments="/path/to/experiment-parent-folder",
    import_images=False,
)

marker = batch.data["GFAP"]
print(type(marker).__name__)
print(marker.df.head())

Direct construction is mainly for advanced code and tests:

from PyFLASH import Attribute

attribute = Attribute("Behavior", behavior_df, experiment)

Antibody, cellMarker, and objectMarker constructors also take an experiment object and a color. The color is optional for Antibody (color=None) but required for cellMarker and objectMarker. objectMarker can take a threshold argument.

Important Attributes¶

Attribute	Applies To	Meaning
`name`	All marker objects	Marker or table name used as the key in `.data`.
`df`	All marker objects	Cleaned `pandas.DataFrame` containing the imported measurements.
`experiment`	All marker objects	Parent experiment. This is removed during pickling and re-linked after loading.
`color`	`Antibody`, `cellMarker`, `objectMarker`	Marker color used by plotting helpers.
`threshold`	`objectMarker`	Colocalisation threshold used when adding colocalisation columns.

Important table columns vary by source file. Common normalized columns include AnimalName, Condition, Region, ROI, and marker-prefixed measurement columns such as GFAP_Count or GFAP_IntDen.

Common Methods¶

Method	Applies To	Use
`set_df(new_df)`	`Antibody` and subclasses	Replace the stored DataFrame and return it.
`addColocData(threshold)`	`objectMarker` through `Antibody` implementation	Add colocalisation count columns from raw colocalisation fields.
`analyse_roi(roi, points, visualise=False)`	`Antibody` and subclasses	Compute distances from points to an ROI line; optionally save a diagnostic figure.
`find_distance_to_ventricle(rois)`	`cellMarker`, `objectMarker`	Add distance-to-ventricle values when ROI data is available.
`find_closest_distances_between_markers(other_marker)`	`cellMarker`, `objectMarker`	Add nearest-neighbor distance and closest-marker columns between markers.

Most lower-level cleaning methods are implementation details. Prefer reading or filtering .df unless you are extending PyFLASH.

Accepted By¶

Marker objects are not usually passed directly to public plotting functions. Instead, pass the parent Batch, Experiment, or DataFrameExperiment and refer to marker tables by name:

from PyFLASH.plotting import plot_histograms

plot_histograms(
    batch,
    marker="GFAP",
    x_attr="GFAP_Area",
    save=False,
)

Internally, marker-aware plots and pipelines look up marker tables in obj.data.

Returned By¶

Marker objects are created by:

Experiment.importCSVs()
Experiment.processData()
MiniExperiment.importCSVs() for flat CSV Attribute tables
Batch.processData(), which merges experiment .data dictionaries into batch.data
load_state, when loading a saved object that contains marker tables

Examples¶

List marker tables:

print(sorted(batch.data))

for name, table in batch.data.items():
    print(name, type(table).__name__, table.df.shape)

Inspect one table:

gfap = batch.data["GFAP"].df
print(gfap.columns.tolist())
print(gfap[["AnimalName", "Condition"]].head())

Filter marker rows before custom analysis:

gfap = batch.data["GFAP"].df
control = gfap.loc[gfap["Condition"] == "Control"].copy()

Use table data from a DataFrameExperiment:

from PyFLASH import from_dataframe

experiment = from_dataframe(
    summary_df,
    group_col="Diagnosis",
    subject_col="Subject",
    data={"Cells": cell_df},
)

print(experiment.data["Cells"].df.head())

Notes¶

Attribute cleans column names and ensures a Condition column based on animal names or condition labels.
Antibody prefixes measurement columns with the marker name.
cellMarker and objectMarker adjust coordinate columns when ROI properties are available.
objectMarker adds colocalisation count-style columns from raw colocalisation measurements.
Source CSV headers are normalized. If a raw column seems missing, inspect marker.df.columns.