Skip to content

What Is Probability (Factor Assignment Confidence)?

Every spot, pixel, or cell in a factor layer is assigned to one or more factors, each with a probability representing the model's confidence (from 0 to 1) that this specific unit belongs to that factor.

What do K1/P1, K2/P2, and K3/P3 represent?

Most factor layers store up to three ranked assignments per spatial unit to capture mixed signals:

  • K1 / P1: The top-ranked factor for this unit, and its probability. Shown as topK / topP in the Inspector.
  • K2 / P2: The second most likely factor, and its probability.
  • K3 / P3: The third most likely factor, and its probability.

These values populate the Factor Probability Table in the Inspector Drawer when selecting individual spots or cells.

How should you interpret the probabilities?

  • High Confidence: If P1 is close to 1.0 (and much larger than P2), the assignment is highly confident — the spatial unit clearly belongs to a single factor.
  • Mixed/Transition State: If P1 and P2 are close in value (e.g. 0.4 and 0.35), the unit represents a mixed signal, transition zone, or boundary between two tissue domains.

Related Pages

Concepts

Manual Pages

How-to Guides