Automated 3D cytoplasm segmentation in soft X-ray tomography

Abstract

Cells' structure is key to understanding cellular function, diagnostics, and therapy development. Soft X-ray tomography (SXT) is a unique tool to image cellular structure without fixation or labeling at high spatial resolution and throughput. Fast acquisition times increase demand for accelerated image analysis, like segmentation. Currently, segmenting cellular structures is done manually and is a major bottleneck in the SXT data analysis. This paper introduces ACSeg, an automated 3D cytoplasm segmentation model. ACSeg is generated using semi-automated labels and 3D U-Net and is trained on 43 SXT tomograms of immune T cells, rapidly converging to high-accuracy segmentation, therefore reducing time and labor. Furthermore, adding only 6 SXT tomograms of other cell types diversifies the model, showing potential for optimal experimental design. ACSeg successfully segmented unseen tomograms and is published on Biomedisa, enabling high-throughput analysis of cell volume and structure of cytoplasm in diverse cell types.

Keywords: Artificial intelligence; Cell biology.

Graphical abstract

Figure 1

Automated cytoplasm segmentation workflow with semi-automated labeling

Figure 2

Dice coefficient according to alteration of the number of the training data

Figure 3

ACSeg segmentation and morphological quantification of T cells The volume (A) and surface area to volume ratio (B) of the cell were measured in the ground truth and prediction of cytoplasm segmentation with the ACSeg. ns ∶ p > 0.05, ∗∶ 0.01 < p<=0.05, ∗∗∶ 0.001 < p<=0.01, using paired t test. N = 10. (C) 3D rendering of Biomedisa (C1) and prediction of the ACSeg (C2), respectively, visual comparison of Biomedisa and our model’s prediction (C3). The box in C3 denotes the area shown up closely in C4.

Figure 4

3D quantification of cytoplasm segmentation accuracy for classical approaches

Figure 5

2D quantification of cytoplasm segmentation accuracy in comparison to SAM

Figure 6

Prediction of the ACSeg over T, BV-2, Huh-7, and MEF cells