Embracing Variability for Better Science – A Must-Attend Seminar at AALAS 

The Digital In Vivo Alliance is excited to invite you to an insightful seminar titled “Embracing Variability to Ensure Reproducibility, Validity, and Translation,” at the upcoming American Association for Laboratory Animal Sciences (AALAS) meeting in Nashville. This session, led by esteemed experts in the field, aims to transform how we perceive variability in preclinical research and its influence on scientific advancement. Attendees will gain valuable insights through discussions on practical, solution-oriented approaches to enhancing reproducibility, validity, and translational relevance in preclinical studies. Don’t miss this opportunity to explore innovative strategies for improving research outcomes! 

Here’s a sneak peak into what you can expect from the session, being led by Dr. Claire Hankenson and Dr. Joe Garner: 

Session Overview: Embracing Variability to Ensure Reproducibility, Validity, and Translation
November 5, 2024  
8 – 10 am Room 101A 

Individual variation is the fundamental essence of life. Without it there would be no evolution and no biology. Human clinical research has long focused on individual variation as the source of knowledge; and modern medicine is pivoting from one-size-never-fits-all interventions to highly personalized biomarker-based interventions. Despite this progress in the clinical space, preclinical researchers frequently aim to minimize variability rather than embracing it. Yet, it’s crucial to acknowledge that individual variability cannot be controlled, and it holds immense value as a subject of study. A growing body of evidence shows that the poor reproducibility of preclinical research, and the poorer forward translation of animal results to humans (i.e. external validity) stems from a lack of tractable biological variation in experimental paradigms. This seminar aims to address some of the multifactorial intricacies that limit reproducibility, to stimulate dialogue, and propose innovative strategies for enhancing the relevance and translatability of preclinical research findings to clinical applications. This seminar will present considerations and discussion around various solution-based, practical approaches such as: advocating for variability and embracing systematic heterogeneity; using home cage digital measures to assess variability in animal behavior within and between cages as well as across sites; inclusion of both sexes and addressing misconceptions and barriers to sex inclusive research; utilizing naturally existing genetic variation in experimental design; and promoting practices to improve generalizability (i.e. effective comparisons between animals and humans), including suggestions for improving rigor and transparency in animal research, encompassing experimental design, statistical assessment, and reporting factors to ensure broader applicability of comparative outcomes. This seminar is for researchers, veterinarians, data scientists, support and care staff, vivarium managers, welfare scientists, ethics committee members, and those interested in improving research reproducibility and the translational relevance of animal studies.  

Seminar Topics and Speakers: 

  • Leaving Stepford Behind: Why We Should, and How We Can, Combine the Best of Human and Animal Experimental Design 

Presenter: Joseph Garner, Professor of Comparative Medicine, and by courtesy, of Psychiatry and Behavioral Sciences, Stanford Medicine  

  • Advocating for Generalizability: Recognizing What Can Confound Expected Research Outcomes 

Presenter: Claire Hankenson, Professor of Pathobiology and Associate Vice Provost for Research and Executive Director of University Laboratory Animal Resources, University of Pennsylvania 

  • Measuring the Variability We Know Exists: Using Innovative Home Cage Digital Measures to Assess Sources of Experimental Variability Including Individual, Cage, and Genetic Effects, and Implications for Study Design and Statistical Analysis 

Presenter: Michael Saul, Computational Scientist, The Jackson Laboratory 

  • Turning Knowing, Into Doing: Evaluation of a Workshop to Train Scientists How to Appropriately Include Sex as a Biological Variable  

Presenter: Brianna Gaskill, 3Rs Senior Scientist, Novartis Institutes for BioMedical Research (NIBR) 

Session sponsored in part by The Jackson Laboratory 

Join us as our presenters guide you through the nuanced world of variability, offering practical solutions that enhance reproducibility, validity and translation. Don’t miss out on this opportunity to elevate your research practices—come, be part of a dialogue that promises to inspire innovation and foster advancements that lead to better science. We look forward to seeing you there! 

Natalie Bratcher-Petersen
Manager of Digital In Vivo Alliance

/0 Comments/by

DIVA to Present at the 75th Annual AALAS Meeting on Enhancing Animal Studies with Digital Biomarkers 

Next week, the Digital In Vivo Alliance (DIVA) is thrilled to take part in the 75th Annual Meeting of the American Association for Laboratory Animal Science (AALAS), a prominent gathering of professionals in laboratory animal science. This event attracts thousands of researchers, veterinarians, and industry leaders dedicated to advancing animal care, research, and management, making it a perfect platform for DIVA to share its work in digital biomarkers. 

DIVA was established by a team of scientists focused on the transformative potential of digital sensors, artificial intelligence (AI), and machine learning (ML) in animal research. Over the past few years, DIVA has worked toward a unified strategic direction, nominated key digital measures that could benefit drug developers, and designed studies to create and validate ML-based algorithms. This collaboration has resulted in a robust validation framework and tangible outcomes we are excited to discuss with the broader research community. 

Here’s a sneak peek into what you can expect from the session, being led by DIVA Chair, Dr. Brian Berridge: 

Session Overview: Leveraging ML-Based Digital Biomarkers in Animal Research 
November 6, 2024 
8 – 10 am Room 207 

Drug development ideally moves seamlessly from basic research to preclinical models and eventually to human studies. Today, the use of digital wearables in clinical settings is reshaping our understanding of human health and disease, offering precise insights and real-time monitoring. DIVA’s work in preclinical digital biomarkers aims to bring similar advancements to animal studies by using home cage-based monitoring and analysis. By applying these digital measures, we hope to improve translational relevance, reproducibility, generalizability, and promote the 3Rs (Replacement, Reduction, Refinement) impact in preclinical research. 

DIVA’s multi-disciplinary collaboration advances digital biomarkers through a comprehensive approach, from development and validation to regulatory acceptance. Our studies include models of drug-induced neuroactivity, Amyotrophic Lateral Sclerosis, Friedreich’s Ataxia, and Dravet Syndrome, allowing us to refine digital biomarkers of activity, locomotion, and seizure. These efforts contribute to training and validating ML algorithms that enhance study reproducibility across sites and settings. This session will share our findings through proof-of-concept vignettes that highlight the advantages of digital biomarkers and DIVA’s unique approach. 

Our seminar will be valuable for researchers, veterinarians, data scientists, vivarium managers, and anyone interested in how digital biomarkers can enhance the reproducibility, generalizability, and ethical standards of animal studies. Below is an outline of the seminar topics and speakers: 

DIVA Seminar Topics and Speakers: 

  • Advancing Animal Research Through Innovation: Building Confidence in Computer-Vision-Derived Digital Biomarkers 
    Presenter: Vivek Kumar, Associate Professor, Center for Addiction Biology, The Jackson Laboratory 
  • Enhancing In Vivo Toxicology Studies with ML-Informed Digital Biomarkers for Improved Sensitivity and Translation 
    Presenter: Brian Berridge, DIVA Chair 
  • Addressing Challenges in Preclinical Neuroscience Research through Home Cage Digital Measures 
    Presenter: Jenny Leedy, In Vivo Scientist, BioMarin 
  • Understanding Cross-Site Reproducibility Challenges in Home Cage Digital Measures 
    Presenter: Michael Saul, Computational Scientist, The Jackson Laboratory 

Sessions sponsored in part by The Jackson Laboratory 

This seminar underscores DIVA’s commitment to harnessing the power of digital measures for robust, reliable, and ethical animal research. We look forward to an engaging session and hope to inspire more researchers and stakeholders to explore the potential of digital measures in their work. We hope we’ll see you there! 

Natalie Bratcher-Petersen
Manager of Digital In Vivo Alliance

/0 Comments/by

Harnessing AI and Digital Sensors for Dynamic Biological Insights: A Reflection on DIVA’s Progress

Like many, I’m excited at the opportunities offered by rapid advancements in digital sensors and AI. I’m particularly interested in the potential for digitalization and AI-based approaches to advance our understanding, modeling, and monitoring complex biology in animals and people. I’ve often thought that our usual time-bound and static approaches to monitoring biological processes fall short in providing continuous and dynamic insights into the way biology (and disease) actually happens- i.e. continuously and dynamically. On the other hand, I have a healthy skepticism of “shiny new rocks” and prefer a strategic approach to engaging them over surfing the wave of excitement they often bring.

This excitement and opportunity was the backdrop for the Digital In Vivo Alliance’s (DIVA) recent annual face-to-face meeting in Redwood City, CA. You won’t be surprised to learn that many of the faces were in the room where we met while others were virtually present via Zoom. Though often less-than-satisfying for participants, those who attended virtually were engaged, actively participated, and presented. This meeting was the third such event for DIVA since our launch nearly 2 years ago and was particularly meaningful since we’re at an important inflection point in our efforts. It was a significant success from my perspective.

DIVA was launched as a collaboration of scientists with mutual interest in leveraging rapid advances in digital sensors, artificial intelligence (AI), and machine learning (ML) to improve the way we do animal studies. We’ve spent the last couple of years aligning on our strategic intent, nominating digital measures that we think will be useful for drug developers, designing and executing studies to support ML-algorithm development, designing a validation framework, and beginning to share the outcomes of our efforts. Our meeting was an opportunity to reflect on our progress, recognize our learnings, and look forward to a DIVA 2.0. 

It was clear that we’ve made significant progress in advancing our foundational portfolio of digital measures represented by the data-rich presentations we heard. That portfolio includes algorithms that enable us to detect and track individual mice with increasing accuracy and efficiency; screen for seizure-related behaviors; measure respiration; and segment activity at levels of resolution that will allow us to quantify locomotory movement, infer sleep, and a host of activity patterns that we’ve yet to explore. We’re applying a rigorous validation protocol to these measures and are exploring novel ways of representing and analyzing these unique measures. We are clearly on the bleeding edge of developing a valuable analytical capability that could transform the way we do animal studies.

Despite our laurels, we recognized that those laurels must add value by addressing fundamental challenges in drug development to support decisions to continue to invest in DIVA and the technologies supporting our digital measures. Accordingly, we spent time talking about what unique values our digital measures might bring that could improve clinical translation, bring process efficiencies, and optimize animal welfare. I think the values are there.  It’s time for us to articulate them and support them with evidence from the studies we’ve executed. 

Lastly, we discussed how to organizationally evolve DIVA to maintain our momentum.  For real impact and to become a standard of practice, we’ll need to recruit and engage a much broader stakeholder group. To maintain and speed our progress, we’ll need to develop more strategic collaborations and leverage a broader set of resources. Transformative innovation doesn’t generally come from individuals but from collaborative teams of people with a common interest and vision. I think we have that in DIVA.             

As a veterinarian, I think we owe it to the research animals we study to maximize the benefit of that study to advance both human and animal health.  As a scientist, I think we leave useful information on the table if we don’t collect as much data as we can from the studies we conduct.  As a drug developer, I think we can do better than we do in translating what we learn preclinically to clinical success. As a potential patient, I have a vested interest in every opportunity to improve our ability to develop safe and effective therapies. As the Chair of DIVA, I welcome the opportunity to continue to work with, support, and learn from this dedicated and creative group of scientists to accomplish all these things.

DIVA is an exemplar for how to collaboratively and strategically engage, develop, and apply novel technologies to solve contemporary problems and support the important work of biomedical science. We’ve established our foundation and built momentum. We’ve clearly identified the problems we want to solve and have guided developers to build solutions that align to those problems. I’m confident that we’ll maintain and even speed our progress throughout the remainder of this year and those to come. 

I’m fortunate to be a member of this team and contributor to this effort. I hope they let me stay. 😊

Brian Berridge, DVM, PhD, DACVP

DIVA Chair

/0 Comments/by