SymbiQuant

SymbiQuant quantifies endosymbiont cells in microscopy images using deep neural network object detection and segmentation to provide polyploid-independent estimates of endosymbiont population size.

Key Features:

• Object detection and segmentation: Uses deep neural networks to detect and segment individual endosymbiont cells within micrographs.
• Polyploid-independent quantification: Produces per-host-cell counts and population size estimates that address limitations of qPCR in polyploid endosymbionts and complement flow cytometry.
• Human-curated output: Supports human curation of detection results to validate and correct automated counts.
• Trained on model system: Initially trained on annotated microscopy images from the aphid/Buchnera endosymbiosis and captures Buchnera population dynamics and phenotypic characteristics over aphid postembryonic development.
• Adaptable training data: Allows replacement of annotated training images to retrain models for other host–endosymbiont systems.

Scientific Applications:

• Population quantification: Provides accurate cell-level counts for estimating endosymbiont population sizes across hosts and tissues.
• Developmental dynamics: Enables measurement of Buchnera population dynamics and phenotypic changes across aphid postembryonic development.
• Comparative endosymbiosis studies: Can be retrained with annotated images from other systems to support organismal, genomic, and tissue- or cell-level analyses of different host–endosymbiont interactions.

Methodology:

SymbiQuant employs deep neural networks for object detection and segmentation trained on annotated microscopy images.