Open datasets can support virtual science education

As science classes move online, some teachers are turning to the Allen Institute’s public resources to build virtual experiments and prepare students for real-world research

September 15, 2020

NoneThis image is a compilation of two openly available resources available from the Allen Institute, the Aging, Dementia and Traumatic Brain Injury Study (left) and the Integrated Mitotic Stem Cell (right), that are being used by educators to support their virtual classroom curriculum.

For the nearly 17 million U.S. college students and their teachers who recently started back to school, everything about this fall has been a scramble.

Colleges and universities are making tough choices between in-person and online learning, and students are sometimes coming back to campus only to find everything about the college experience to be different, or they are choosing to stay distant from school entirely. This year’s strangeness is backdropped against the novel coronavirus pandemic, which has seen more than 88,000 cases on college and university campuses since last spring.

For those schools that have moved education entirely or partially online, science classes pose their own set of challenges — teachers used to leading lab courses with hands-on experiments have had to scramble to find replacements. Some are turning to public datasets to supplement their coursework.

We talked with three college neuroscience and biology professors who are using public Allen Institute resources — data, lesson plans, and other tools — in their virtual classrooms this fall. Two neuroscience professors are using datasets from the Allen Brain Map, an online portal with publicly available data and other resources generated by researchers in the Allen Institute for Brain Science, a division of the Allen Institute. One biology professor is using both the neuroscience datasets and a lesson plan and resources about cell division from the Allen Institute for Cell Science to supplement her introductory biology course.

Elizabeth Glater, Ph.D., a neuroscientist and associate professor at Pomona College in Claremont, California, was teaching a course on genes and behavior when the pandemic hit the U.S. last spring. By the time her school shifted to virtual classes, she’d gotten through two out of three planned lab modules.

Glater had the idea that her students could do computational analyses with online datasets instead, but there was not enough time to pull a virtual module together. Their spring semester ended in mid-May, and the next day Glater and her fellow science teachers started planning for a virtual lab experience for the fall, anticipating that classes might remain online — which they did for her school.

“It takes a lot to pull together all the information and make it clear and interesting and engaging,” Glater said. “It’s been a real process.” 

She and her colleagues found a lot of websites with exercises analyzing imaginary data that weren’t quite right, Glater said. She wanted her students to have a realistic research experience, even if laboratory experiments weren’t possible. 

Built for scientists, adapted for education

NoneNeuroscientist and associate professor Elizabeth Glater (top right), co-instructors and students examine public online data from the Allen Institute’s Aging, Dementia and Traumatic Brain Injury study.

Glater was aware of the Allen Institute’s public datasets from her neuroscience research, but it hadn’t occurred to her to use them for teaching until the pandemic. Allen Brain Map, the web portal, was designed with researchers in mind, but teachers have adapted some of the datasets for educational use over the years, said Allen Institute Training and Outreach Specialist Kaitlyn Casimo, Ph.D. Casimo has been working with Allen Institute neuroscientists and cell biologists to develop free online education resources for high school and college-level biology classes.

“When I started working on our educational resources, we knew some educators were using our data in their classes, but we didn't have anything specifically to support them or get them started,” Casimo said. “We have a massive amount of data, so students can dig in to generate a novel research question and tackle it using data we've made available online. Or they can combine our resources with other data they've collected in class.”

The real-world applications of the Allen Institute datasets suited Glater’s purposes perfectly. She wants her students to experience what actual neuroscientists experience, which these days means working with large datasets, including those generated by others. She’s using both the Allen Cell Types Database, which includes data about the genes switched on in individual brain cells from mouse and human brains, as well as data from the Aging, Dementia and Traumatic Brain Injury Study, in her fall introductory neuroscience class.

“This idea of doing an experiment and getting the data and everything is done in your own lab – that’s an old-fashioned way of doing science,” Glater said. “I want my students to be aware of what else is out there. Even if you’re still doing experiments in the lab, these big datasets can inform what you do.” 

Introducing students to modern neuroscience

Ashley Juavinett, Ph.D., a neuroscientist and assistant professor at the University of California San Diego, has been using Allen Institute resources in her classes even before the move to virtual teaching. As a graduate student, she used some of the neuroscience data for her own research, and attended the Summer Workshop on the Dynamic Brain, a two-week computational neuroscience course run by the Allen Institute and the University of Washington.  

"For many students, our Allen Institute labs are the first time they’ve seen this kind of data. It’s an introduction to this very cutting-edge world of neuroscience,” Juavinett said.

When Juavinett started her position as a professor two years ago, she incorporated several of those datasets into her courses. She teaches a neurobiology lab course that is (normally) mostly hands-on, but also includes some data analysis. She also teaches a neural data science class that teaches coding and data analysis; in that class, students rely on open-access datasets for independent research projects.

NoneThis screenshot from Ashley Juavinett's website show three different lesson plans she developed for interacting with Allen Institute datasets.

Juavinett has already run both of these classes remotely; UCSD is on the quarter system and moved classes online at the start of spring quarter, and she taught one class in the summer quarter as well. Like Glater, she chose Allen Institute datasets in part for her students to get a real-world experience with modern neuroscience and data analysis. And the public datasets also let her introduce concepts and virtual experiments that aren’t possible to do in a teaching lab — even when students can attend in person.

“There are certain things we simply can’t do in the lab. For example, you might want to start students out doing really simple electrophysiology, where they record from one neuron or one nerve cord. But neuroscience today is more like recording from 1,000 neurons at the same time,” Juavinett said. “For many students, our Allen Institute labs are the first time they’ve seen this kind of data. It’s an introduction to this very cutting-edge world of neuroscience.”

Challenges to remote science education

One of the biggest challenges for Juavinett’s students is not unique to science – she thinks the lack of in-person social interactions is one of the worst parts about remote learning. Students are missing the chance to immediately talk with their classmates about concepts they’re having trouble with, and it’s also harder for students to build a connection and trust with their teachers, she said. 

“I can't really be present with them in the same way online, and that’s hard,” Juavinett said. “As teachers, we have to really think about what the goal of our class is and what we want our students to learn. How do we get at that online?”

Sarah Latchney, Ph.D., a neuroscientist and assistant professor at St. Mary’s College of Maryland in St. Mary’s City, is teaching a hybrid biology class this fall – she’ll be in the classroom with some of her students, and her classes will be streamed simultaneously to those students who are attending remotely. She’s thinking about how to make sure her remote students get the same quality of education as her in-person students – especially in light of barriers to science education that pre-date the pandemic. Lower-income students and underrepresented minority groups already face barriers to accessing STEM fields for jobs and education, Latchney said, and she’s worried about how these students will fare when they are also missing out on in-person instruction.

“We’re trying the best that we can to retain our students and engage them,” said Latchney, who’s planning to use the Allen Institute for Cell Science’s lesson plan and the Integrated Mitotic Stem Cell, an online visualization tool about cell division, for her fall introduction to biology class. “That’s why I’m excited about these open access resources on the Allen Institute website because I really think open-access science helps to minimize some barriers.”

Glater said she thinks her students might have challenges with their preconceived notions about learning science – will they feel like they’re learning how to do research if they can’t do hands-on experiments? She thinks reframing the important aspects of science for them will help.

“When we think of a lab, the first thing that comes to mind is the physical space where you do experiments. But I think we need to reassure the students that there are more important things to learn from lab than just the technical skills,” Glater said. “We want them to learn how to think, how to plan experiments, how to analyze data. Those are the main things I want my students to learn this year.”

Get the latest news from the Allen Institute.

more news

Subscribe to newslette