Determining growth rates from bright-field images of budding cells through identifying overlaps

Abstract

Much of biochemical regulation ultimately controls growth rate, particularly in microbes. Although time-lapse microscopy visualises cells, determining their growth rates is challenging, particularly for those that divide asymmetrically, like Saccharomyces cerevisiae, because cells often overlap in images. Here we present the Birth Annotator for Budding Yeast (BABY), an algorithm to determine single-cell growth rates from label-free images. Using a convolutional neural network, BABY resolves overlaps through separating cells by size and assigns buds to mothers by identifying bud necks. BABY uses machine learning to track cells and determine lineages and estimates growth rates as the rates of change of volumes. Using BABY and a microfluidic device, we show that bud growth is likely first sizer- then timer-controlled, that the nuclear concentration of Sfp1, a regulator of ribosome biogenesis, varies before the growth rate does, and that growth rate can be used for real-time control. By estimating single-cell growth rates and so fitness, BABY should generate much biological insight.

Shugoshin maintains mitotic arrest in response to improper kinetochore-microtubule interactions independent of the Bub1-pH2A axis

“…Timelapse assays were performed as described (Leontiou et al, 2022). Cell outlines were segmented using the baby algorithm (Pietsch et al, 2023). To quantify the kinetochore retention time of GFP-Bub1 fluorescence, we create a projection of the maximum values from all GFP sections, then divide the median fluorescence of the brightest 5 pixels within each cell by the median fluorescence of the cell as a whole.…”

Deep learning-driven imaging of cell division and cell growth across an entire eukaryotic life cycle

“…Imaging the full yeast life cycle revealed significant morphological and optical diversity during sporulation ( Fig 2 A ), germination ( Fig 2 B ) and mating ( Fig 2 C ). Current yeast detection algorithms, such as YeastNET 58 , YeaZ 55 , BABY 68 , Cellpose 60 , and the yeast compatible cell AC/DC functionalities 69 , however, do not specifically segment cells according to life cycle stage, and often depend on pixel intensity thresholds that can change due to morphological and optical heterogeneity. To overcome this challenge, we leveraged solutions for pixel intensity- and morphology-independent image segmentation developed for human silhouette recognition 70 and bacterial and mammalian cell biology 71, 72 based on the openpose and cellpose architectures.…”