Electron microscopy (EM) enables high-resolution visualization of protein distributions in biological tissues. For detection, gold nanoparticles are typically used as an electron-dense marker for immunohistochemically labeled proteins. Manual annotation of gold particle labels is laborious and time consuming, as gold particle counts can exceed 100,000 across hundreds of image segments to obtain conclusive data sets. To automate this process, we developed Gold Digger, a software tool that uses a modified pix2pix deep learning network capable of detecting and annotating colloidal gold particles in biological EM images obtained from both freeze-fracture replicas and plastic sections prepared with the post-embedding method. Gold Digger performs at near-human-level accuracy, can handle large images, and includes a user-friendly tool with a graphical interface for proof reading outputs by users. Manual error correction also helps for continued re-training of the network to improve annotation accuracy over time. Gold Digger thus enables rapid high-throughput analysis of immunogold-labeled EM data and is freely available to the research community.

No abstract

Electron microscopy (EM) enables high-resolution visualization of protein distributions in biological tissues. For detection, gold nanoparticles are typically used as an electron-dense marker for immunohistochemically labeled proteins. Manual annotation of gold particle labels is laborious and time consuming, as gold particle counts can exceed 100,000 across hundreds of image segments to obtain conclusive data sets. To automate this process, we developed Gold Digger, a software tool that uses a modified pix2pix deep learning network capable of detecting and annotating colloidal gold particles in biological EM images obtained from both freeze-fracture replicas and plastic sections prepared with the post-embedding method. Gold Digger performs at near-human-level accuracy, can handle large images, and includes a user-friendly tool with a graphical interface for proof reading outputs by users. Manual error correction also helps for continued re-training of the network to improve annotation accuracy over time. Gold Digger thus enables rapid high-throughput analysis of immunogold-labeled EM data and is freely available to the research community.

Integral membrane proteins such as ion channels, transporters, and receptors shape cell activity and mediate cell-to-cell communication in the brain. The distribution, quantity, and clustering arrangement of those proteins contribute to the physiological properties of the cell; therefore, precise quantification of their state can be used to gain insight into cellular function. Using a highly sensitive immunoelectron microscopy technique called sodium dodecyl sulfate-digested freeze-fracture replica immunogold labeling (SDS-FRL), multiple membrane proteins can be tagged with different sizes of immunogold particles at once and visualized two-dimensionally. For quantification, gold particles in the images must be annotated, and then different mathematical and statistical methods must be applied to characterize the distribution states of proteins of interest. To perform such analyses in a user-friendly manner, we developed a program with a simple graphical user interface called Gold In-and-Out (GIO), which integrates several classical and novel analysis methods for immunogold labeled replicas into one self-contained package. GIO takes an input of particle coordinates, then allows users to implement analysis methods such as nearest neighbor distance (NND) and particle clustering. The program not only performs the selected analysis but also automatically compares the results of the real distribution to a random distribution of the same number of particles on the membrane region of interest. In addition to classical approaches for analyzing protein distribution, GIO includes new tools to analyze the positional bias of a target protein relative to a morphological landmark such as dendritic spines, and can also be applied for synaptic protein analysis. Gold Rippler provides a normalized metric of particle density that is resistant to differences in labeling efficiency among samples, while Gold Star is useful for quantifying distances between a protein and landmark. This package aims to help standardize analysis methods for subcellular and synaptic protein localization with a user-friendly interface while increasing the efficiency of these time-consuming analyses.