Harnessing the power of citizen scientists

Summary

The Rosalind Franklin Institute’s Artificial Intelligence and Informatics (AI&I) theme and imaging teams are working with citizen scientists on a range of projects.


Collaborators

As one of the Franklin’s founding members, we have worked closely with Diamond Light Source from our inception. A commissioning call through the Diamond eBIC facility enabled first time community access to key components of our cryo-electron tomography pipeline.

diamond.ac.uk

Working with researchers from Diamond Light Source and University of Oxford, the Franklin team is using cutting edge imaging techniques to identify nanoparticles and classify nanoparticle catalysts. Nanoparticles are incredibly small, measuring between 1 and 100 nanometres – one thousandth of a human hair. They are used in everything from the manufacture of medicines, flatscreen TVs and agrochemicals to water purification. The research could lead to more efficient and environmentally friendly nanoparticle catalyst systems.

Dr Michele Darrow, the Franklin’s Head of Data Strategy for Cryo-Electron Imaging, said: ‘We hope that by collecting basic information about the number of individual nanoparticles, their sizes and how many of them are part of clusters we can better understand their role and make them more efficient and environmentallyfriendly when they are used in agriculture and other industries.’

Using Scanning Transmission Electron Microscopes (STEM), Dr Darrow’s colleagues and their counterparts at Oxford and Diamond Light Source have taken thousands of incredibly detailed images of these tiny particles.

Although AI found and categorised many of the nanoparticles, it struggled to separate others from the background of the image and mis-labelled some clusters.

To address this, they have combined the best in machine learning with human classification – which is where the citizen scientists come in.

Send in the Zooniverse

The partners have found willing volunteer collaborators by launching the nanoparticle catalyst project on the Zooniverse. The Zooniverse is a citizen science portal with 1.6 million registered users ranging from school children to retired people. They are currently hosting more than 50 projects.

It is based on the idea that anyone with a computer and a bit of time on their hands can help move research forward. Volunteer citizen scientists help researchers to do science that otherwise would not be possible. The Zooniverse was established by the University of Oxford, Adler Planetarium and the Citizen Science Alliance – with many projects supported by the Alfred P. Sloan Foundation.

Indeed, some of those citizen scientists have already helped Dr Darrow with several health-focused projects posted on her Zooniverse Science Scribbler pages – including studying images of cells affected by the neurological disorder, Huntington’s disease and visualising cells as they are infected by a virus, Virus Factory.

Citizen scientists on the Zooniverse are reinforcing the work of Dr Darrow, her colleagues and AI in an ongoing project, Placenta Profiles, which aims to explore in detail the complex structure of the placenta, starting by finding all of the mitochondria (the energy producing factories inside of cells).

Dr Darrow said: ‘How well the placenta works is related to its development and structure but that’s hard to study because there are lots of different cell types, each playing different roles and taking different forms. Placenta Profiles is giving professionals and citizen scientists the opportunity to do important work analysing the complex structure of the placenta that we can then feed back to machine learning, making it faster and better in the future.’

The nanoparticles project mentioned above, named Key2Cat, has already sparked great interest amongst the citizen scientists. Within a week of data being posted in May 2022, hundreds of Zooniverse members identified and marked around 170,000 nanoparticles and clusters on the STEM images.

The researchers used these marks to create a new dataset showing each individual nanoparticle. The team put together a few questions and asked for the help of citizen scientists again to classify the nanoparticle catalysts, before the results are again analysed by AI. The last citizen science leg of the project is ongoing now – click here to help out!

Providing the citizen scientists with the tools for the job

The Franklin is also working with the University of Oxford and University College London to develop and test out software tools so citizen scientists on the Zooniverse can add the human touch to AI analysis of large datasets. The topics range from astrophysics to ecology and the activity of proteins in human cells.

‘With the software tools being developed we’re looking for the lower abundance items that are not obvious. We’re looking for the uncommon. We’re really looking for the anomalies, the things that are missed most of the time.’
Head of Data Strategy for Cryo Electron Imaging

Oxford astrophysicists (and citizen scientists) will be the first to test out the new tools as their datasets are ready for analysis. They and the citizen scientists will analyse masses of data about objects at the farthest reaches of the Universe.

Dr Darrow, who received her PhD for imaging work around the structural basis of protein misfolding disorders, will use the tools – and the support of citizen scientists – to focus again on proteins inside the cell. For the first time, advanced imaging equipment is providing us with large numbers of high quality volumes showing proteins engaging in their normal activities within human cells.

Dr Darrow explains: ‘Biochemists have been able to look at the structure of a purified protein outside the cell environment for almost 100 years but when you take the protein out of its natural environment its structure can change, so sometimes you don’t get an accurate picture.’

Using Science Scribbler projects on the Zooniverse, she will again enlist citizen scientists in the study of biological data – the basic science that could eventually identify opportunities for researching potential treatments for disease.

More case studies