Frequently asked questions
The basics of what we’re building and why.
What is mZero?
mZero is a global scientific effort to turn mosquito behavior data into training data. Our models learn which kinds of compounds are most likely to change a mosquito’s decision to land on human skin.
What is the key question mZero is attempting to answer?
How do you change a mosquito's decision to land on human skin?
Who can join mZero?
University labs, research institutes, and qualified independent researchers running behavioral assays on mosquitoes.
How do I apply?
Please request access here. If your lab is accepted, we'll follow-up with onboarding instructions.
What will mZero provide?
We will provide each accepted lab with compounds, standardized assay materials, and access to shared behavioral data. In some cases, we'll fund a postdoctoral researcher.
What problem are you solving?
Malaria is resurgent, with more than 250 million cases per year. It's one of the most technically and morally challenging problems of our time.
Who are some of your partners?
Partnering with leaders in malaria research, machine learning, and compute, including the University of Toronto’s Department of Chemical Engineering & Applied Chemistry, the United States Department of Agriculture (USDA), the NVIDIA Inception Program, Ifakara Health Institute, Lambda, the Google for Startups Cloud Program, and Spectrum Impact.
Why is this effort even more relevant now?
Decades can pass with little progress, then months can produce decades’ worth of progress. The acceleration of machine intelligence, together with the global scientific community’s commitment to generating the data needed to unlock its potential, gives us a rare opportunity to make significant progress in the near-term.
What is the contact/non-contact assay?
Please see the full assay protocol and videos of the assay here.
Why screen in Aedes if the goal is malaria?
Aedes aegypti Liverpool is the most widely maintained and best-characterized mosquito colony in the world and allows for standardization of assay results.
Can I use mZero data in my own research?
Yes, all data is free to use, cite, and build on under CC BY 4.0, a standard open license.
What paper topics could a participating lab pursue?
- Inter-lab reproducibility of the contact/non-contact assay across geographies and mosquito strains.
- Cross-genus concordance of landing inhibition between Aedes aegypti and Anopheles gambiae.
- Benchmarking in-silico repellency predictions against prospective assay outcomes.
- Structure–activity relationships in a new scaffold class identified through the open atlas.
- Mode-of-action studies linking repellent chemistry to mosquito odorant receptor interactions.
How does the predictive model work?
The model strengthens patterns associated with compounds that actually work and weakens patterns associated with compounds that fail (which is the case for most compounds). Across many comparable experiments, it becomes better at recommending the next experiment: which compound, dose, formulation, or controlled assay adjustment is most likely to produce useful evidence and increase the probability of finding another effective compound. The learning signal strengthens, and we take a step closer to finding compounds that will reduce malaria transmission.
How will our data be used?
Data from mZero will be used to train our AI models with the goal of predicting effective compounds. Wet lab compound efficacy data closes the learning loop for a model, even if some noise remains. The model strengthens patterns associated with compounds that actually work and weakens patterns associated with compounds that fail. Across many comparable experiments, it becomes better at ranking which compound to test next, increasing the probability of finding another effective compound.
How is data quality controlled across labs?
All labs run a fixed protocol on the same assay, pair every treated run with a same-day solvent control, and test a shared compound panel at the same concentrations.
We will report the following metrics to assess data reliability:
- Inter-assay correlation (r)
- Hit re-test rate
- Inter-lab variance (CV)
- z-score