MyVivarium: A cloud-based lab animal colony management application with realtime ambient sensing

Abstract

Management of research-animal colonies is vital for lab productivity in preclinical research. Labs often rely on inefficient paper-based or spreadsheet-based methods to manage animal colonies. Dedicated software-based solutions are generally expensive and many lack remote access. Currently available open-source alternatives are difficult to implement and deploy. These solutions also do not have the capability to track ambient variables that affect colony well-being. We built MyVivarium, an open-source database management web application to address these gaps. MyVivarium can be easily deployed to the cloud and sustained for a cost comparable to starting and maintaining a lab website. Using MyVivarium, lab members can collaboratively track individual animals within a database. Physical identities of cages map onto the database using QR codes, enabling quick and easy record-keeping on mobile devices. Lab administrators can assign tasks to users with reminders for experiments or cage maintenance. Finally, we designed a low-cost system to sense ambient humidity, temperature, vivarium worker activity, and room illuminance. These data are then sent to MyVivarium in realtime providing information relevant to colony well-being. Taken together, MyVivarium is a novel, open-source, cloud-based application template that provides a low-cost, simple, and efficient way to digitally manage research-animal colonies.

Keywords: Biological database; IoT sensors; Mouse facility; Open-source technology.

Graphical abstract

Fig. 1

MyVivarium System Architecture Overview. MyVivarium system architecture consists of three main components: the client side, the server side, and external services. Each component interacts with the others to provide a robust and secure web application for managing vivarium operations. Client-side: MyVivarium's front-end is built using standard web technologies, including HTML, CSS, and JavaScript, ensuring that the user interface is interactive and responsive. Users can interact with the application through web browsers such as Chrome and Firefox with cross-device compatibility. External services: Cloudflare provides DNS services and security features, including DDoS protection, to enhance the reliability and security of the MyVivarium application. Server-Side: The backend is built on a LAMP stack, consisting of Linux (operating system), Apache (webserver), MySQL (database), and PHP (scripting language). The database includes tables such as users, breeding, holding, mouse, files, settings, litter, tasks, notes, and strains, which store various types of data required for the application's functionality. IoT Data: The system integrates IoT sensor data to monitor ambient parameters such as temperature, humidity, illuminance, and presence. Scheduler: A scheduler (cron job) handles tasks such as sending reminder emails utilizing the email_queue for processing. Cron jobs automate database queries and sending emails through SMTP using PHPMailer.

Fig. 2

MyVivarium web application navigation overview. The MyVivarium web application navigation is intuitive, structured into “Home”, “Dashboards”, and “Settings”. The “Home” section includes “Cages Summary”, “Tasks Summary”, and “General Notes”. The “User Login” provides links for registration and password recovery. “Dashboards” gives access to holding cage and breeding cage management, including adding, editing, viewing, and deleting cages, printing cage cards, maintenance logs, and realtime ambient sensing via the IoT sensor datastream. “Settings” includes user profile management, task management, and admin options such as managing users, lab details, mouse strains, IACUC protocols, and data export in CSV format. The Logout option allows users to log out securely. This structure ensures easy access to all functionalities, with specific options available based on user roles and permissions.

Fig. 3

MyVivarium web application usage metrics for a single lab. a. Number of active users accessing the Sathyanesan Lab MyVivarium across different pages. b. Views per active user across page titles. c. Engagement time (in minutes) for a specific page. In panels a-c, note the highest metrics for holding-cage-related tasks since these are cages in higher number and are frequently used in experiments. Task management is also used significantly by users. d. MyVivarium usage percentage distribution by operating system, e. Device category, and f. Browser. In panels d-f, note the diversity in distribution across different categories, indicating pan-OS, pan-device, and pan-browser usability.

Fig. 4

Task management ensures adherence to weaning and experimental timelines. a. Eight weaning timelines for which tasks were created for efficient management at different times (green upside-down triangle; Task Creation Date) between litter date of birth (blue circle, Litter DOB) and actual wean date (red diamond; Actual Wean Date). b. Eight experimental timelines for which tasks were created (green upside-down triangle; Task Creation Date) also at different dates across the timeline. Dashed line in panels a and b denote the target day (day 0) for weaning (in a) or a timed experiment (in b). Note that in a, as inferred from the task completion date all litters are weaned on the target wean date. In b, all experimental tasks are completed by target date.

Fig. 5

Design, assembly, and deployment of MyVivarium RPi-IoT sensor system. a. Components of Pi-IoT sensor system include 1: Jumper wires (to extend reach) 2: SHT30 Temperature and humidity sensor with plastic shell casing 3: VL53L1X distance or time-of-flight sensor which can measure distance to a person (vivarium-worker) 4: VEML7700 illuminance sensor which can detect lights ON and OFF 5 and 6: STEMMA-QT connectors (for Raspberry Pi GPIO pins to sensors and between sensors) 7: Raspberry Pi 4B (2 GB) in a case with a cooling fan. b. Fritzing sketch of serial connections between I²C sensors and Raspberry Pi GPIO pins based on a validated sensor module build. Wire color designations are as follows – red: power (+ve), black: ground (-ve), yellow: SCL (serial clock), blue: SDA (serial data). Note that in a, the color of SDA (data) is white. c. Assembled sensor module deployed in our vivarium room. Sensors have been affixed to the lower part of the mouse cage transfer station. The Raspberry Pi has been placed in a corner of a mouse cage rack.

Fig. 6

Continuous realtime ambient factor recording over a period of months using MyVivarium RPi-IoT sensor system. a. Temperature (ºC) recorded over four months in vivarium rooms 1 (blue) and 2 (orange). Temperature in Dayton, OH is plotted as reference (grey). Dotted lines represent minimum (“min” = 18 ºC) and maximum (“max” = 26 ºC) recommended temperatures. Time-window (grey shaded rectangle) between arrows (12 midnight May 2nd, 2024 – 12 midnight May 3rd, 2024) is shown in greater detail in b. c. Relative humidity (%) recorded over four months. Humidity in Dayton, OH is plotted as reference (grey). Dotted lines represent minimum (“min” = 30 %) and maximum (“max” = 70 %). Time-window (grey shaded rectangle) is shown in greater detail in d. e. Illuminance sensor values recorded over months. Asterisk indicates an abnormal event detected by the sensor due to an erroneous manual override of the automated light cycle f. Sensor value change over the course of 24 hours shows vivarium lights OFF (light bulb turned off) and lights ON (light bulb turned on) in rooms 1 and 2 set to a 6 AM – 6 PM light cycle. g. Proximity sensor values for rooms 1 and 2 recorded for months h. Vivarium worker presence or activity is detected as negative deflections. Minimum and maximum temperature and humidity thresholds are based on suggested recommendations by The Guide (Office of Laboratory Animal Welfare, National Institutes of Health) and the Jackson Laboratory , .

Fig. 7

User notification of sensor readings and automated alerts. a. Increased humidity in vivarium rooms 1 (blue) and 2 (orange) on February 27th, 2024. b. General user notification using the “Sticky notes” on MyVivarium. c. Abnormal illuminance event and automated alert testing (arrow indicating when black unreflective paper was used to block the illuminance sensor from light exposure) shows reduced illuminance (orange) at 3:30 PM on March 25th d. Alert message triggered by Grafana and relayed via a Google chat webhook app to a dedicated Google Space.