Myers Leverages CRISPR Technology in First-of-its-Kind Systematic Mapping Project

With the rise of services like 23andme and Ancestry.com, individuals have more information about their genetics than ever before. As more people participate in these services, researchers can study genetic data and start putting together the puzzle that explains why people do and don’t get certain diseases. There is not a one-to-one relationship between a gene and a disease; it is a complex array of factors and relationships between genes that today’s scientists are just starting to untangle.

Department of Computer Science & Engineering Professor Chad Myers has been working on decoding this complex network of genomic data for over a decade. His lab focuses on understanding gene function and leverages computational biology to make inferences about biological networks.

“Surprisingly with all the information scientists have gathered, machine learning models still don’t perform anywhere near as well as they should to predict who will actually get a disease,” Myers said. “We can get the genes, we can build models, but we are still not able to explain why these diseases are inherited in most cases. That is the basic problem we are working on - can we build better models that use genetic data to predict and reach their full potential?”

The Myers Lab has a wide range of projects looking at genes on a macro and micro scale. While his group continues to leverage artificial intelligence (AI) and machine learning methods to analyze combinations of genes using data collected from the larger human population, he is also working to understand basic genetic relationships in a single cell line with the help of CRISPR technology. CRISPR (short for “clustered regularly interspaced short palindromic repeats”) is a technology that research scientists use to selectively modify the DNA of living organisms.

“The CRISPR cell-line project is a systematic mapping project. CRISPR allows us to take a human cell line, introduce combinations of mutations into genes, and measure how much it affects their ability to grow in a dish. We are mapping a systematically surveyed network of these interactions in an exhaustive way so that we can start to generalize some rules around interactions between genes.”

The groundwork for this project began over a decade ago in a joint effort with his experimental collaborators at the University of Toronto. The project focused on genetic interactions in yeast cells. In 2016, Myers and Toronto’s Charlie Boone and Brenda Andrews published the first complete genetic interaction network for any species. The introduction of CRISPR technology opened the doors to expand this process to human cells.

“Very few people are doing systematic surveys of genetic interactions; that is what is unique about our project. In our yeast research, we mapped a social network of genes. So when we do these experiments, we are connecting genes that when deleted or mutated together, have a surprising effect on growth. Once you start analyzing that network, you can build a map of how everything is related. We are now applying that methodology to the human genome. Mapping these networks is a way to learn about how genes function, which can broadly push forward our knowledge of human biology and serve as the basis for other research endeavors.”

Using CRISPR technology, Myers’ experimental collaborators are able to conduct large-scale genetic screens in a controlled setting. In one set of dishes, CRISPR can test 18,000 combinations of mutations with a single gene over a two-week period. For each gene, multiple CRISPR guides cut the genome in three to four different places in different cells to average the observations in that gene. If all the cut areas have some mutations, they can be confident that the gene is important for growth.

“One surprising thing we learned in yeast is that out of 6,000 genes, only 1,000 or so are essential for growth. The same concept holds true in the human genome. Biological systems are extremely robust and can function and grow even with a number of mutations. There is a lot of interest in understanding how a system can be so robust in a pretty severe environment.”

The exploratory and systematic nature of this project requires a true team effort. After the success of the yeast project in 2016, Myers, Boone, and Andrews teamed up with Jason Moffat, an expert in human functional genomics at the Sick Kids Hospital in Toronto, to help transition from yeast to human cells. The team works together to design experiments and interpret the resulting data. The Toronto group carries out the large-scale CRISPR screens while the Myers Lab handles the interpretation of the huge amounts of sequencing data, which is used to understand the effects of the mutations introduced in each screen.

“My lab has developed new computational approaches to measure effects of gene mutations from CRISPR screen sequencing data and ensure we are measuring the right things. It is a complicated process. We have made a lot of progress in understanding the details of how CRISPR technology works in this setting and the types of biases and experimental artifacts that show up as patterns in the resulting data. Data processing and statistical normalization approaches are critical to ensure we are capturing the correct information, measuring fitness and cell growth.”

The experimental and computational methods developed throughout the project will have lasting impacts on a broad range of genomic research efforts. The Myers Lab has published the methods they used to extract data from these CRISPR experiments, including how to score and normalize data, and understand patterns. Along the way, they have published many papers detailing the insights and learnings of applying machine learning and data mining to this unique type of data.

“That has been a key output of this work - understanding how to use CRISPR technology to get the measurements that we want and then interpreting them. This has been a long, iterative process to build a road map for this new technology in order for future researchers to generate tons of data that we can interpret reliably.”

In addition to the methods used in this multi-year project, the final dataset produced from the CRISPR findings will also serve as a launching pad for future genetic research. Similar to their work with yeast, this is a new type of network dataset that no one has ever published before, and can be integrated with a wealth of other human genomic data produced over the past two decades.

“Once we understand the principles for how genes interact using this cancer cell line in the lab, we and other researchers can use our dataset to do more research within real human populations. Each person has around one million mutations that uniquely distinguish them, which makes tracking down the effects of individual or combinations of mutations tricky. Starting with a cell line in which we can introduce two and only two mutations gives us a solid foundation to learn from. We are learning the principles of how and when genes interact with a hope to apply those learnings to other settings, like interpreting the variation that shows up in each of our genomes.”

This systematic survey of gene networks will also help researchers better understand the functions of genes, specific disease variants, and combinations of mutations that might impact phenotypes in humans.

“We are mapping the edges of a gene network, which will help us better define relationships between genes and start discovering how everything is connected. We can take a gene that we know little about, identify that it is clustered with these other genes, and use the commonalities in that cluster to get a better understanding of the gene in question.”

The CRISPR project is a marquee example of the impact bioinformatics and computational biology (BICB) can have across the scientific community. Myers has served as the co-director of graduate studies for the BICB program at the University of Minnesota since 2017, and helps students with backgrounds in computer science or biological sciences blend their skill sets to tackle some of the world’s most pressing issues.

“Computing as a basis for biomedical science is not going away; it is only going to grow in its role. All of these fields can benefit from the developments in machine learning and AI. Programs like BICB include a community of students with research questions spanning a wide range of biological areas. They get to learn from each other in a unique way. I’m excited to see what kind of impact that will have at the University of Minnesota and research beyond the university setting. I’m excited to continue to teach at this intersection and mentor younger faculty as this field grows.”

In spring 2024, Myers was named a Distinguished McKnight University Professor, and received the Award for Outstanding Contributions to Graduate and Professional Education for his work with the BICB program and beyond.

“In the computational biology space, there is a huge need for teachers. At the University of Minnesota, there’s an amazing medical school and a variety of biology disciplines from plant science to developmental biology to drug discovery. There is a demand for people in all of these areas to have computational skills to enable their research. Through BICB, we have taught several hundred students and given them their first exposure to using computers to solve problems and answer questions about biological data. That’s a very rewarding step to see students take.”

Learn more about the Myers Lab and the CRISPR project on the lab website.

Share