Normalization is the process of correcting for systematic sources of variation across wells and plates, enabling meaningful comparison of all compounds within a screen. Two primary strategies are used in uHTS analysis.

Plate-Based Normalization

In plate-based normalization, each compound’s measurement is normalized using the control wells located on the same plate. This is the standard approach for uHTS and corrects for plate-to-plate variability in signal magnitude, which can arise from:

• Small differences in target dispense between plate batches.

• Slight variations in incubation time or temperature between plates.

• Gradual changes in instrument performance over the course of a multi-day screen.

By calculating Robust Z scores using each plate’s own neutral reference population, plate-based normalization ensures that a compound’s score reflects its binding activity relative to the assay conditions on that specific plate, rather than being confounded by plate-to-plate drift.

Recommendation: Plate-based normalization is the recommended default for Dianthus (uHTS) experiments. It requires a sufficient number of neutral reference wells per plate (typically a full column or row) to calculate a stable reference distribution.

Screen-Based (Global) Normalization

Screen-based normalization pools the reference information from all plates across the entire screen to calculate a global reference distribution. Compound scores are then calculated relative to this global reference.

This approach is most appropriate when:

• The number of reference wells per plate is too low to calculate a reliable per-plate distribution.

• The assay signal is exceptionally stable across all plates (confirmed by very consistent per-plate Z’ values).

Caution: Screen-based normalization assumes that assay conditions are uniform across all plates. If systematic between-plate variation exists (e.g., a shift in labeling reagent concentration mid-screen, changes in excitation power), global normalization will introduce systematic biases into compound scores and may inflate or deflate hit rates for entire plate batches.

Alternative Normalization Methods

For assays with spatial artefacts (e.g., edge effects or column-to-column drift), more advanced methods are available:

These methods are not typically required for a well-optimized Dianthus uHTS assay, but may be applied to salvage data from plates with documented spatial artefacts.