Quantification of Osteoclasts in Culture, Powered by Machine Learning

Abstract

In vitro osteoclastogenesis is a central assay in bone biology to study the effect of genetic and pharmacologic cues on the differentiation of bone resorbing osteoclasts. To date, identification of TRAP+ multinucleated cells and measurements of osteoclast number and surface rely on a manual tracing requiring specially trained lab personnel. This task is tedious, time-consuming, and prone to operator bias. Here, we propose to replace this laborious manual task with a completely automatic process using algorithms developed for computer vision. To this end, we manually annotated full cultures by contouring each cell, and trained a machine learning algorithm to detect and classify cells into preosteoclast (TRAP+ cells with 1–2 nuclei), osteoclast type I (cells with more than 3 nuclei and less than 15 nuclei), and osteoclast type II (cells with more than 15 nuclei). The training usually requires thousands of annotated samples and we developed an approach to minimize this requirement. Our novel strategy was to train the algorithm by working at “patch-level” instead of on the full culture, thus amplifying by >20-fold the number of patches to train on. To assess the accuracy of our algorithm, we asked whether our model measures osteoclast number and area at least as well as any two trained human annotators. The results indicated that for osteoclast type I cells, our new model achieves a Pearson correlation (r) of 0.916 to 0.951 with human annotators in the estimation of osteoclast number, and 0.773 to 0.879 for estimating the osteoclast area. Because the correlation between 3 different trained annotators ranged between 0.948 and 0.958 for the cell count and between 0.915 and 0.936 for the area, we can conclude that our trained model is in good agreement with trained lab personnel, with a correlation that is similar to inter-annotator correlation. Automation of osteoclast culture quantification is a useful labor-saving and unbiased technique, and we suggest that a similar machine-learning approach may prove beneficial for other morphometrical analyses.

Biomanufacturing in low Earth orbit for regenerative medicine

“…Currently, tools are being developed to automate terrestrial cell culture ( Cohen-Karlik et al., 2021 ) and biofabrication methods ( De Pieri et al., 2021 ). These technologies could be applied to research in LEO, allowing experiments to run autonomously and potentially scale in progression from research to clinical applications ( Vieira et al., 2021 ).…”

“…Additionally, research groups are applying ML and AI to improve cellular products ( Cohen-Karlik et al., 2021 ; Mota et al., 2021 ), biomaterial manufacturing ( An et al., 2021 ; Lee et al., 2020 ), and disease modeling ( Severson et al., 2021 ). Utilizing existing datasets, both from terrestrial experiments as well as LEO-based experiments ( da Silveira et al., 2020 ), ML approaches could be built into the automated LEO platforms.…”

OC_Finder: Osteoclast Segmentation, Counting, and Classification Using Watershed and Deep Learning

“…There are two recent related works (Cohen-Karlik et al, 2021;Emmanuel et al, 2021) that reported software to detect osteoclasts. The foremost important difference to note is that these two works did not release the datasets they used and their software to the public.…”

OC_Finder: A deep learning-based software for osteoclast segmentation, counting, and classification

“…There are two recent related works 17,18 that developed software to detect osteoclasts. The foremost important difference to note is that these two works did not release the datasets they used and their software to the public.…”

“…In addition, each of them has notable differences from the current work. The work by Cohen-Karlik et al 17 used a different neural-network framework to detect cells and classify osteoclasts. Their network outputs bounding boxes of cells while OC_Finder segments the cell region boundaries.…”