Tag Archive for: computational

Mihailo Kaplarevic

Extracting actionable research data faster, with fewer hassles

Mihailo Kaplarevic

Mihailo Kaplarevic, Ph.D., the newly minted Chief Research Information Officer at Children’s National Hospital and Bioinformatics Division Chief at Children’s National Research Institute, will provide computational support, advice, informational guidance, expertise in big data and data analyses for researchers and clinicians.

Kaplarevic’s new job is much like the role he played most recently at the National Heart, Lung and Blood Institute (NHLBI), assembling a team of researchers and scientists skilled in computing and statistical analyses to assist as in-house experts for other researchers and scientists.

NHLBI was the first institute within the National Institutes of Health (NIH) family to set up a scientific information office. During his tenure, a half-dozen other NIH institutions followed, setting up the same entity to help bridge the enormous gap between basic and clinical science and everything related to IT.

“There is a difference compared with traditional IT support at Children’s National – which will remain in place and still do the same sort of things they have been doing so far,” he says of The Bear Institute for Health Innovation. “The difference is this office has experience in research because every single one of us was a researcher at a certain point in our career: We are published. We applied for grants. We lived the life of a typical scientist. On top of that, we’re coming from the computational world. That helps us bridge the gaps between research and clinical worlds and IT.”

Ultimately, he aims to foster groundbreaking science by recognizing the potential to enhance research projects by bringing expertise acquired over his career and powerful computing tools to help teams achieve their goals in a less expensive and more efficient way.

“I have lived the life of a typical scientist. I know exactly how painful and frustrating it can be to want to do something quickly and efficiently but be slowed by technological barriers,” he adds.

As just one example, his office will design the high-performance computing cluster for the hospital to help teams extract more useful clinical and research data with fewer headaches.

Right now, the hospital has three independent clinical systems storing patient data; all serve a different purpose. (And there are also a couple of research information systems, also used for different purposes.) Since databases are his expertise, he will be involved in consolidating data resources, finding the best way to infuse the project with the bigger-picture mission – especially for translational science – and creating meaningful, actionable reports.

“It’s not only about running fewer queries,” he explains. “One needs to know how to design the right question. One needs to know how to design that question in a way that the systems could understand. And, once you get the data back, it’s a big set of things that you need to further filter and carefully shape. Only then will you get the essence that has clinical or scientific value. It’s a long process.”

As he was introduced during a Children’s National Research Institute faculty meeting in late-September 2019, Kaplarevic joked that his move away from pure computer science into a health care and clinical research domain was triggered by his parents: “When my mom would introduce me, she would say ‘My son is a doctor, but not the kind of doctor who helps other people.’ ”

Some of that know-how will play out by applying tools and methodology to analyze big data to pluck out the wheat (useful data) from the chaff in an efficient and useful way. On projects that involve leveraging cloud computing for storing massive amounts of data, it could entail analyzing the data wisely to reduce its size when it comes back from the cloud – when the real storage costs come in. “You can save a lot of money by being smart about how you analyze data,” he says.

While he expects his first few months will be spent getting the lay of the land, understanding research project portfolios, key principal investigators and the pediatric hospital’s biggest users in the computational domain, he has ambitious longer-term goals.

“Three years from now, I would like this institution to say that the researchers are feeling confident that their research is not affected by limitations related to computer science in general. I would like this place to become a very attractive environment for up-and-coming researchers as well as for established researchers because we are offering cutting-edge technological efficiencies; we are following the trends; we are a secure place; and we foster science in the best possible way by making computational services accessible, affordable and reliable.”

tube labeled "CRISPR"

$2M from NIH to extract meaningful data from CRISPR screens

tube labeled "CRISPR"

Protein-coding genes comprise a mere 1% of DNA. While the other 99% of DNA was once derided as “junk,” it has become increasingly apparent that some non-coding genes enable essential cellular functions.

Wei Li, Ph.D., a principal investigator in the Center for Genetic Medicine Research at Children’s National in Washington, D.C., proposes to develop statistical and computational methods that sidestep existing hurdles that currently complicate genome-wide CRISPR/Cas9 screening. The National Institutes of Health has granted him $2.23 million in funding over five years to facilitate the systematic study of genes, non-coding elements and genetic interactions in various biological systems and disease types.

Right now, a large volume of screening data resides in the public domain, however it is difficult to compare data that is stored in one library with data stored at a different library. Over the course of the five-year project, Li aims to:

  • Improve functional gene identification from CRISPR screens.
  • Develop new analyses algorithms for screens targeting non-coding elements.
  • Study genetic interactions from CRISPR screens targeting gene pairs.

Ultimately, Li’s work will examine a range of disease types. Take cancer.

“There is abundant information already available in the public domain, like the Project Achilles  from the Broad Institute. However, no one is looking to see what is going in inside these tumors,” Li says. “Cancer is a disease of uncontrolled cell growth that makes tumors grow faster.”

Li and colleagues are going to ask which genes control this process by looking at genes that hit the brakes on cell growth as well as genes that pump the gas.

“You knock out one gene and then look: Does the cell grow faster or does it grow more slowly? If the cell grows more slowly, you know you are knocking out a gene that has the potential to stop tumor growth. If cells are growing faster, you know that you’re hitting genes that suppress cancer cell growth.”

In a nutshell, CRISPR (clustered regularly interspaced short palindromic repeats) screens knock out different genes and monitor changes in corresponding cell populations. When CRISPR first became popular, Li decided he wanted to do something with the technology. So, as a Postdoc at Harvard, he developed comprehensive computational algorithms for functional screens using CRISPR/Cas9.

To reach as many people as possible, he offered that MAGeCK/MAGeCK-VISPR software free to as many researchers as possible, providing source code and offering internet tutorials.

“So far, I think there are quite a lot of people using this. There have been more than 40,000 software downloads,” he adds. “It’s really exciting and revolutionary technology and, eventually, we hope the outcomes also will be exciting. We hope to find something really helpful for cancer patients.”

Research reported in this publication was supported by the National Human Genome Research Institute of the National Institutes of Health under award number R01HG010753.