A subscription to JoVE is required to view this content. Sign in or start your free trial.
The imaging platform "The Lifespan Machine" automates the lifelong observation of large populations. We show the steps required to perform lifespan, stress resistance, pathogenesis, and behavioral aging assays. The quality and scope of the data allow researchers to study interventions in aging despite the presence of biological and environmental variation.
Genetically identical animals kept in a constant environment display a wide distribution of lifespans, reflecting a large non-genetic, stochastic aspect to aging conserved across all organisms studied. This stochastic component means that in order to understand aging and identify successful interventions that extend the lifespan or improve health, researchers must monitor large populations of experimental animals simultaneously. Traditional manual death scoring limits the throughput and scale required for large-scale hypothesis testing, leading to the development of automated methods for high-throughput lifespan assays. The Lifespan Machine (LSM) is a high-throughput imaging platform that combines modified flatbed scanners with custom image processing and data validation software for the life-long tracking of nematodes. The platform constitutes a major technical advance by generating highly temporally resolved lifespan data from large populations of animals at an unprecedented scale and at a statistical precision and accuracy equal to manual assays performed by experienced researchers. Recently, the LSM has been further developed to quantify the behavioral and morphological changes observed during aging and relate them to lifespan. Here, we describe how to plan, run, and analyze an automated lifespan experiment using the LSM. We further highlight the critical steps required for the successful collection of behavioral data and high-quality survival curves.
Aging is a complex, multifaceted process characterized by a decline in the physiological function of an organism, which leads to an increase in the risk of disease and death over time1. Lifespan, measured as the time from birth or the onset of adulthood until death, provides an unambiguous outcome of aging2 and an indirect but rigorously quantitative proxy for measuring the relative rate of aging between populations3. Aging studies often depend on accurate measurements of lifespan, similar to clinical trials, to compare outcomes between one population exposed to an intervention and an unexposed control group. Unfortunately, reproducibility issues pervade aging research, sometimes due to statistically underpowered experiments4 and often because of the inherent sensitivity of lifespan assays to subtle variations in the environment5. Robust experiments require multiple replicates of large populations, and this process particularly benefits from the experimental scalability offered by automation6.
The rigorous demands of lifespan assays originate from the unpredictability of the aging process itself. Isogenic individuals housed in identical environments display different death times and rates of physiological decline7, suggesting that lifespan involves a high degree of stochasticity7,8. Therefore, large populations are required to measure quantitative changes in the aging process, such as changes in the mean or maximum lifespan, and to overcome biases arising from individual variability. In addition, a capacity for high-throughput lifespan assays is crucial to support studies of survival curve shapes and models of the dynamics of aging9.
The nematode Caenorhabditis elegans is an invaluable model for aging research due to its short lifespan, genetic tractability, and rapid generation time, which underscore its suitability for high-throughput aging and lifespan assays. Traditionally, the lifespan in C. elegans has been measured by following a synchronized, small population of about 50-100 animals over time on solid media and writing down the time of individual deaths. As animals age and lose mobility, manually scoring the death times requires individually prodding the animals and checking for small movements of the head or tail. This is usually a tedious and laborious process, though efforts have been made to accelerate it10,11,12. Importantly, slow experimental pipelines hinder progress in our understanding of aging and the effectiveness of tested interventions.
To meet the demands of aging research for quantitative data, many technologies have been developed for automating data collection, including a remarkable range of approaches from microfluidic chambers to flatbed scanners13,14,15,16,17,18. The LSM differs from other methods in its extensive optimization for the collection of highly precise and accurate lifespan data, which is achieved through the development of careful equipment calibration protocols combined with an extensive software suite that allows users to validate, correct, and refine automated analyses13. Though the software can, in principle, be applied to diverse imaging modalities, in practice, most users use flatbed scanners modified to allow for fine-tuned control over the environmental temperature and humidity - factors of critical importance due to their major effect on lifespan19. The LSM takes images of nematodes every 20 min over intervals ranging from days to months, depending on the environmental conditions and genotype. The data produced are of much higher temporal resolution compared to data from manual assays, and the images collected provide a permanent visual record of the nematode position across the lifespan. Using machine-learning methods, death times are automatically assigned to each individual. These results can be rapidly, manually validated using a client software called "Worm Browser". As a result of its hardware and software, the LSM can generate survival curves that are statistically indistinguishable from manual death scoring at the hands of experienced researchers, with the added advantage of decreased workload and higher scalability13.
The latest version of the LSM also allows for the study of behavioral aging by collecting morphological and behavioral data throughout the nematode's life and reporting it along with the lifespan of each individual. In particular, the LSM captures the time of each animal's vigorous movement cessation (VMC), a landmark often used to quantify the "healthspan" of an individual as distinct from its lifespan. By simultaneously collecting lifespan and behavioral aging data, the LSM supports the study of interventions that may have differential effects on different phenotypic outcomes of aging20. A variety of macroscopically observable phenotypes can be used to study behavioral aging, such as body movement or pharyngeal pumping21, tissue integrity22, and movement speed or stimulus-induced turning17. Comparisons between different aging phenotypes can support analyses of the causal structure of aging processes. For instance, the comparison between VMC and lifespan was recently used to characterize two distinct aging processes in C. elegans23.
While initially developed to measure lifespan in C. elegans, the LSM supports the collection of survival and behavioral data from a range of nematode species, including C. briggsae, C. tropicalis, C. japonica, C. brenneri, and P. pacificus23. The technology facilitates the study of the effect of biological and environmental interventions on lifespan, stress resistance, and pathogen resistance and can be coupled to experimental tools such as targeted assays of RNA interference or auxin-inducible protein degradation systems. To date, it has been used in the scientific literature for a wide range of applications6,24,25,26,27,28,29,30.
Here, we outline a step-by-step protocol for performing a Lifespan Machine experiment using agar plates, from the initial stages of the experimental setup to the output of the resulting survival curves. A distinctive feature of the LSM is that the effort is highly front-loaded, meaning that the majority of the researcher's time is spent during experimental setup and, to a small degree, during post-image acquisition. The data collection is completely automated for the whole duration of the experiment and allows the researcher to have a "hands-free" experience. The steps described here are held in common among many different types of survival assays - the same experimental setup is performed for lifespan, thermotolerance, oxidative stress, and pathogenesis assays. In the representative results section, we discuss a subset of data from a recently published manuscript to illustrate the effectiveness of the analysis pipeline and highlight the most important steps during image analysis23.
1. Software and hardware requirements
Supplementary Figure 1: Lifespan Machine hardware. One flatbed scanner unit with an open lid to show the loaded plates, which are placed facing down into 16 openings cut on a rubber mat. The rubber mat is placed on the surface of a glass scanner. Labels for the conditions are written on the sides of the plates to avoid issues during image analysis. Marking tape with the number ("1") and/or name of the device ("Jabba") facilitates later verification of the sample location when working with multiple scanner devices. More details about the LSM hardware components are found elsewhere13. Please click here to download this File.
2. Setup prior to the day of the experiment
3. Setup on the day of the experiment
4. Pre-image acquisition
NOTE: A comprehensive flowchart summarizing all the software-based steps during image acquisition is depicted in Figure 1.
Figure 1: Graphical overview of the Lifespan Machine image analysis pipeline. The pre-, during, and post-image acquisition steps are largely performed on the web interface (WI, in red) and on the Worm Browser (WB, in green). Some steps are performed in other platforms (O, in blue), such as TXT documents in step 3a, Photoshop or equivalent in step 4b, and JMP or equivalent in step 13. Please click here to view a larger version of this figure.
Figure 2: Preview capture image and scan area selection. (A) For each scanner in the experiment, a preview capture image is generated. (B) Selection of one row of plates at a time (red boxes), which increases the speed of scanning and prevents worm motion blur as a result of scanning areas that are too wide. Please click here to view a larger version of this figure.
5. Image acquisition
NOTE: The following steps can be performed both while the experiment is running or after it has finished.
Figure 3: Specification of the plate locations for each scanner using sample masks. To ensure the independent analysis of plates within the column selections shown in Figure 1, individual plates must be identified by generating an image mask composite. (A) A capture of the scans of the scanners is opened with an image manipulation software (note the name of the scanner "han" above a scanned selection, and "a-d" referring to each of the columns). (B) The individual steps of mask generation to mark the location of each plate in the mask composite require the background to be set to black, (C) the removal of jagged edges and edges of non-selected plates by the expanding and then shrinking of the background, and (D) selecting the foreground plates and filling the areas entirely with white pixels. (E) For the LSM to recognize individual plates in the scanned rows, each white region in a row is filled with a different shade of gray, usually in increasing brightness. (F) At this stage, the mask is saved (LZW compression with no layers specified if generated in Photoshop). The mask is then scanned by the Worm Browser, and a visualization of the mask by the software is generated. A correct mask visualization should display one defined square per plate with a small cross at the center and a different color for each row. Please click here to view a larger version of this figure.
Figure 4: Plate quality control using the web interface. Censoring of suboptimal plates on the web interface before the worm movement analysis is crucial for speeding up the image processing pipeline. Examples of plates subject to removal include conditions of (A) desiccation, (B) contamination, or (C) fogging, as opposed. (D) Optimal plates to be included in further analysis. A scale bar of 10 mm is superimposed onto a preview capture image. Please click here to view a larger version of this figure.
6. Post-image acquisition
NOTE: After worm detection is completed, all data collected from the experiment must be aggregated over time to track each individual across their lifespan and identify all the individuals' death times. Wait until all animals in the experiment have died and until all the worm detection jobs have been completed, and then perform the following steps:
Figure 5: Animal storyboard on the Worm Browser. (A) All stationary worms are shown in chronological order of machine-annotated death time. To navigate the storyboard, press the buttons on the (B) bottom-right corner, and (C) save the annotations often. (D) The images with a non-gray background depict two worm death events (early death as green, later death as red), which can either occur when two worms die close to each other, or when dead worms are moved by a passing worm and are, thus, detected as dead twice. (E) A red tag in the bottom corner of an image identifies worms with a detected death time; (F) a green tag indicates where an object did not remain still long enough to record a death time. (G) Multiple worms in the same frame can be flagged by pressing shift and left-clicking. (H) Non-worm objects are excluded from the analysis by a right-click.(I) Exploded worms are censored from the analysis by clicking on the corresponding image (a by-hand annotation window opens) and pressing shift and right-clicking until an "animal exploded" message appears. A scale bar of 0.5 mm and labels are superimposed on the screenshot of a Worm Browser window. Please click here to view a larger version of this figure.
Figure 6: Inspecting objects and annotation of death times on the Worm Browser. Left-clicking on any object on the Worm Browser storyboard opens a new interface and allows the user to inspect the object's movement dynamics. On the right side, the (A) movement score is displayed, which quantifies the object movement; this is estimated by the change in pixel intensities between consecutive observations. Additionally, on the right side, (B) the change in the total object intensity is displayed, which quantifies changes in the object size. On the left side, the upper bar shows the (C) machine estimate of the death time, while the bottom bar is the (D) human by-hand annotation. Clicking on any point of the bars and pressing the space key allows the user to move through the time frames in which the worm has been imaged. On these bars, pink represents the time spent in vigorous movement, red represents the time spent in death, and yellow is everything in between. The time spent in expansion and contraction after the death time is shown in green. Labels are superimposed on the screenshot of a Worm Browser window. Please click here to view a larger version of this figure.
Figure 7: Population summary statistics on the Worm Browser. Population statistics for the scanner device "obiwan", with a plot of the survival (left panel) and a scatter plot of the vigorous movement cessation (VMC) time versus the death time (right panel). The plotted are details of (A) one condition, obtained from (B) one scanner achieved by first selecting (C) the survival grouping by strain. (D) The square shapes in the scatter plot depict the by-hand annotated events, while (E) the circular shapes depict the machine-annotated events. (F) By-hand annotation is often required for death events that occur early or (G) those where the time of vigorous movement cessation time coincides with the death time. Labels are superimposed on the screenshot of a Worm Browser window. Please click here to view a larger version of this figure.
Experimental reproducibility in lifespan assays is challenging and requires both tightly controlled experimental conditions and large populations to achieve sufficient statistical resolution4,36. The LSM is uniquely suitable for surveying large populations of animals in a constant environment with high temporal resolution. To demonstrate the capability of the LSM, highlight the crucial steps of analysis, and help researchers to prioritize their labor efforts, we ...
Here, we provide a detailed, accessible protocol for performing an experiment using the latest version of the Lifespan Machine. We have shown that the critical step for achieving well-resolved survival curves is the manual exclusion of non-worm objects during post-image acquisition. Manual death time annotation has a small effect on the overall shape of the survival curves, demonstrating that fully automated death time estimation is efficient even without manual annotation (Figure 8). On the...
The authors declare that they have no competing interests.
We thank Julian Ceron and Jeremy Vicencio (IDIBELL Barcelona) for producing the rpb-2(cer135) allele. This project was funded by the European Research Council (ERC) under the European Union's Horizon 2020 research and innovation programme (Grant Agreement No. 852201), the Spanish Ministry of Economy, Industry and Competitiveness (MEIC) to the EMBL partnership, the Centro de Excelencia Severo Ochoa (CEX2020-001049-S, MCIN/AEI /10.13039/501100011033), the CERCA Programme/Generalitat de Catalunya, the MEIC Excelencia award BFU2017-88615-P, and an award from the Glenn Foundation for Medical Research.
Name | Company | Catalog Number | Comments |
1-Naphtaleneacetic acid (Auxin) | Sigma | N0640 | Solubilize Auxin in 1M potassium hydroxide and add into molten agar |
5-fluoro-2-deoxyuridine (FUDR) | Sigma | F0503 | 27.5 μg/mL of FUDR was used to eliminate progeny from populations on UV-inactivated bacteria |
Glass cleaner | Kristal-M | QB-KRISTAL-M125ml | |
Hydrophobic anti-fog glass treatment | Rain-X Scheibenreiniger | C. 059140 | |
Rubber matt | Local crafstman | Cut on a high-strength EPDM rubber sheet stock | |
Scanner glass | Local hardware supplier | 9" x 11.5" inch glass sheet | |
Scanner plates | Life Sciences | 351006 | 50 mm x 9 mm, polystyrene petri dish |
USB Reference Thermometer | USB Brando | ULIFE055500 | For calibrating temperature of scanners |
Request permission to reuse the text or figures of this JoVE article
Request PermissionThis article has been published
Video Coming Soon
Copyright © 2025 MyJoVE Corporation. All rights reserved
We use cookies to enhance your experience on our website.
By continuing to use our website or clicking “Continue”, you are agreeing to accept our cookies.