When Diana Murray first started working with two high school students on an online summer cancer systems biology course in 2018, she never imagined that the initiative, now called DREAM-High, would snowball to 56 participants by 2026. “The class started as a consolation prize for high school students who didn’t get selected for a different program funded by the Palazzo Strozzi Foundation,” shares Murray, who in addition to being a computational systems biologist is also the Program Director of Research and Education at Columbia University. And she proceeds to explain.

Not many American academic institutions can boast a connection to Palazzo Strozzi, a 14th century masterpiece of Renaissance architecture and a foundation that supports it, but Columbia University’s Department of Systems Biology is a special case. Andrea Califano, who is a Professor of Chemical and Systems Biology and one of the department cofounders and chairmen, is originally from Italy. “He knows people from the Palazzo Strozzi Foundation USA,” explains Murray. Established in 2010, the Foundation works to introduce American students to Italian culture, arts and history. Every year, the Foundation offers the High School Renaissance Award to 18 students to spend a month in Italy studying the renaissance, the culture and the arts.

About 500 students apply, and those who aren’t selected are understandably disappointed. To offer them an alternative, the program organizers had asked Califano whether Columbia University could create some form of a science program so that students who weren’t chosen to go, could apply for it. Califano said yes.

Back in 2018, Columbia participated in an undergraduate cancer biology course funded by the National Institute of Health’s Cancer Systems Biology Consortium (CSBC). “That undergraduate course was very successful, but then two of my high school students who weren’t chosen to go to Italy took it and they loved it!” Murray says. “So we decided to make this course accessible to other high school students. The undergraduate course was called Mini-DREAM, so we named the high school course DREAM-High.”

Murray worked with three talented students to develop the initial curriculum for the course. Two of them, Palazzo Strozzi high school students, had taken the course previously and the third one, an undergraduate from SUNY Oneonta, who Murray met serendipitously at a math fair in Ithaca, NY, was an expert in a free, open-source programming language R, used for statistical computing and graphical visualization. “She added many novel components that are still key to the program today. We developed this curriculum together in 2020, and we have offered it every summer since,” says Murray. “The only prerequisite for the course is students’ curiosity, so when they apply, they just need to write a paragraph telling why they want to take it and send me their resume. Their science teacher needs to write a two-sentence letter recommending them—and that’s it.”

Since 2020, the DREAM-High course developed into a highly specialized, research-oriented online STEM enrichment program that introduces high school students to modern cancer systems biology, genomics and computational data science. During the course, students delve into topics such as cancer biology, cancer genomics, patient clinical data, gene expression analysis and protein interaction networks, while also learning the computational skills used in modern biomedical research. Rather than focusing on traditional laboratory exercises, the program relies on hands-on analysis of real-world breast cancer datasets and uses the R language and a cloud based RStudio environment to analyze the data. “Students meet together on zoom for three hours once a week, with a choice of two weekly offerings and also explore a different research topic every week asynchronously,” says Murray. These required activities are submitted through the program’s Learning Management System Pathwright.

The course starts with “toy data sets,” notes Murray, so that students can learn basic statistics and data visualization like graphs and heat maps, in which they map numbers to colors and look for patterns in the data. During the second week, students move onto analyzing breast cancer patients’ data. “We use data from the Cancer Genome Atlas, a huge project funded by the NIH that covers 33 different types of cancers, with 35,000 patients contributing data,” says Murray. “We focus on data from 1,082 breast cancer patients. We look into different types of breast cancer and students are stunned to see that there are so many.”

Students also dive into details like the status of estrogen and progesterone receptors (ER and PR) and the receptor HER2 and how they affect cancer growth. The students are especially interested in triple negative breast cancer (TNBC) where breast cells lack all three receptors making them difficult to treat. Murray encourages students to think outside the box. “For example, we ask them ‘can men have breast cancer?’ and most of them will say ‘no.’ But then they look at the data, and they discover that there are indeed some male breast cancer patients.”

Week three is even more exciting as students explore the complexity of real data. “Think of the gene expression data from 1000 patients with 18,000 genes, that's not something they can look at in an Excel document easily,” Murray shares. In an earlier assignment, students examine the data from 70 patients by hand, so this represents a big jump. Now, students must really trust their analysis of the larger data sets, so they revisit the heat maps.

“It's very clear from the breast cancer heat map that a certain set of patients has TNBC because these patients share very similar gene expression signatures,” Murray says. “That kind of freaks the students out, because they look at the clinical data, they look at the gene expression data, then we combine the information, and lo and behold, the pattern that they find from gene expression alone matches the patients with triple negative breast cancer.” That’s the point where data research truly becomes meaningful.

Murray’s favorite part of the course begins right after that. “We move onto human cancer cell lines and how quickly cells move on different substrates,” she says. “We work with two breast cancer cell lines, where one type moves very quickly on the substrate and the other moves slowly.” Not surprisingly, the fast cells are a model for triple negative breast cancer, the more aggressive type; and the ones that move slower are estrogen receptor positive, so they are a model for less aggressive cancer types. Students switch from patients’ data to an experimental system where they can start to understand the TNBC biology.

Students then do a differential gene expression analysis in which they subtract the gene expression values of the cell line that moves slowly from the cell line that moves quickly. “The hypothesis is that the genes that have the highest difference in the fast versus slow cell lines are important for the biology of triple negative breast cancer,” says Murray. “And we show that it’s true.” As students do functional analysis on these genes, they find that the genes are indeed enriched for biological processes like cell migration and other factors related to aggressive cell migration and metastasis. “I love that activity,” says Murray, “because it shows that what patients experience has an underlying biological basis.”

During week five, students do prognostic modeling, where they discover that certain genes are predictive of how long a patient may live. That ties to what oncologists use in the clinic to understand patients’ prognosis. Come week six, students review what they learned and publish their activities on DREAM-High website. During the course, they also have an opportunity to speak with an undergraduate student and a post-doctoral researcher and ask them career-related questions. “They are very comfortable with these junior level scientists, so they ask them about what classes to take and what kind of research to pursue, and what they did to get where they are today,” Murray says. “Senior scientists also join to share their research to demonstrate to students where all of this can lead. It’s an inspiring aha moment for students.”

The program continues to evolve. This year DREAM-High has 56 attendees, a big jump from last year’s 36 and the program’s humble beginnings. The dramatic increases leveraged interest from several other institutions: Stanford University, the Institute for Systems Biology in Seattle and, most recently, Duke University, all of whom support students from nearby high schools to join the program.

There are also several exciting new additions to the program’s curriculum. “This year our students will do a senior project,” says Murray. “It’s a big ask, so we will help them with that.” Another new addition is the introduction of Python, a different programming language that is used more commonly than R. A program alumna, who has been hired as an intern for this year’s class, has migrated the course to Python. “We are going to have them side by side, the activity in R and the activity in Python, and students will match them as much as possible,” says Murray. “It's really exciting to be able to work with both coding languages side by side. R is easier to learn than Python, but Python is what most data scientists use.”

This year, the program will also welcome a patient advocate, Dr. Angie Bowen, a Science Program Manager from Sage Bionetworks. “She is an amazing woman who has been through breast cancer, who understands the science behind the disease, and also works with scientists,” says Murray. “Her role here is to give us all a reality check, to remind us that cancer affects real people, not just abstract numbers. So the research we are doing is ultimately to save people’s lives.” Next year, Murray plans to loop in several artists who have battled cancer and have expressed their cancer experience through art because this will add another human dimension to the data work. “Students love hearing human stories and what these people have gone through.”

More exciting developments loom ahead for DREAM-High. The program was recently awarded a $75,000 grant from a CSBC cross-consortium project proposal, which will allow it to further grow the effort in partnership with other CSBC centers, enroll more students and make it more global and universal. “Next year we are going to create a publication,” Murray shares her vision for DREAM-High. “And we are going to structure this program so that anyone can roll out this course at any academic level. We want this experience to be available not only for high school students, but also for middle school students, college students, medical students and clinicians.” The goal is to create a final product that people can use at different levels.

“We started with just two high school students who wanted to do something different for the summer,” Murray says. “Come next year, the program will go on to have a life of its own. Countless students cite DREAM-High as a major influence in their college and career choices, so it has become a valuable pipeline to exciting state-of-the-art fields related to cancer, systems biology, and computational analysis of biomedical data.”