Season 2 - Episode 20: Harnessing the Power of Data Science in Cancer Research

Listen and Subscribe to Inside Cancer Careers

• Apple Podcasts
• Spotify Podcasts

Episode Guest

 

Jill Barnholtz-Sloan, Ph.D.

As the acting director of CBIIT, Dr. Barnholtz-Sloan is responsible for advancing open data, open software, and open science for NCI. Alongside her peers who make up CBIIT’s Offices and Programs—the Office of the CIO, the Office of Data Sharing, the Informatics and Data Science Program, and the Office of the Director—she continues to engage in ground-breaking efforts in data science, including the Cancer Research Data Commons (CRDC) and the Childhood Cancer Data Initiative (CCDI).

As the Associate Director for Informatics and Data Science Program, she lead efforts at CBIIT to shape informatics and data science strategies and foster collaboration within NCI and across NIH and the cancer research community. Additionally, she is pursuing a robust research agenda in descriptive epidemiology and etiology of brain tumors as an intramural senior investigator in NCI’s Division of Cancer Epidemiology and Genetics (DCEG) Trans-Divisional Research Program. Thus, as both an active researcher and administrator, she has insight into how data can be translated into real-world solutions to help diagnose, prevent, and treat cancer.

Show Notes

• Jill Barnholtz-Sloan, Ph.D. (email:jill.barnholtz-sloan@nih.gov)
• NCI Center for Biomedical Informatics and Information Technology
• NCI Office of Data Sharing
• Childhood Cancer Data Initiative
• NIH Sex and Gender Differences in Cancer Virtual Workshop Series
  • Feb 22, 2024
  • March 28, 2024
  • May 23, 2024
• Cancer Research Data Commons

Ad: NCI Cancer Data Science Training 

Your Turn Recommendations

• “…set boundaries so that you can keep for yourself some kind of balance in your life” from Dr. Barnholtz-Sloan
• 3Blue1Brown - neural network course

Transcript

Oliver Bogler:

Hello, and welcome to Inside Cancer Careers, a podcast from the National Cancer Institute where we explore all the different ways people fight cancer and hear their stories. I'm your host, Oliver Bogler from NCI Center for Cancer Training.

Today, we're talking about the intersection of informatics, data science, and epidemiology with a researcher engaged in these areas who's also taken a leadership role at the NCI.

Listen through to the end of the show to hear our guests make some interesting recommendations and where we invite you to take your turn. And of course, we're always glad to get your feedback on what you hear and suggestions on what you might like us to cover. The show's email is NCIICC@nih.gov.

So it's a pleasure to welcome Dr. Jill Barnholtz-Sloan, Acting Director of the NCI Center for Biomedical Informatics and Information Technology, or CBIIT as we call it, and also Associate Director of Informatics and Data Science. Welcome.

Jill Barnholtz-Sloan:

Thank you so much for having me, Oliver. I'm thrilled to be here.

Oliver Bogler:

Can you describe for us your day-to-day responsibilities in these roles?

Jill Barnholtz-Sloan:

Sure, happy to. So the Center for Biomedical Informatics and Information Technology at the NCI has a mission to empower NCI staff and the cancer research community with data science, information technology, and data sharing tools that they need to advance our knowledge of cancer. And we do that by trying to accelerate groundbreaking research using data and technologies to minimize the burden of cancer.

So the Center for Biomedical Informatics and IT or CBIIT covers really a broad range of things. And so maybe I'll just touch on what kind of the main buckets are of things that we touch. So in the Office of Data Sharing led by Jamie Guidry Auvil, they do everything data sharing. So it's all about understanding the different types of data. What are the rules and regulations to be compliant with the NIH data management and sharing policy? How do they work with NCI staff and investigators both in the intramural and extramural research program to help them on their journey for their research so that their data can then be shared and accessed widely in the cancer research community.

Oliver Bogler:

And why is data sharing important?

Jill Barnholtz-Sloan:

Well, I think data is everywhere, right? And it's coming so fast and furious and it's, it's something that we all need in order to advance even our day-to-day ,sort of who we are and what we're doing and how we live. I think a lot of people don't realize how much data is used around them in their everyday life. We just don't really think of that.

And cancer in particular, which of course is what the National Cancer Institute is focused on, is such a hard problem. We've come a long way in many facets of cancer, but in many other facets we haven't. And so sharing data that investigators are generating and allowing other smart, creative people from all over to use those data hopefully will lead us to those novel insights that we need to really solve all of the problems in cancer that we currently face.

The other big thing that Jamie's team does in the Office of Data Sharing is that they oversee the Childhood Cancer Data Initiative project, which is a really critical project for us to advance our knowledge about pediatric cancer. While pediatric cancer is very rare, many of the different types really have just such a heavy impact on the children and their families, and many of them do not have as long of survival as we would like to see for those children to live full and complete lives. So what they have done there is they've developed protocols and alliances across the country to try to bring into a singular data ecosystem all of this information, all types of information on pediatric cancers, again, so that they can make it publicly available, so that everybody around the world can utilize those data to make impact for children with cancer.

Oliver Bogler:

I guess particularly important in that scenario because thankfully those tumors are rare and so any one center may not have enough data to do anything meaningful, but together they can accomplish things.

Jill Barnholtz-Sloan:

Right, I think for the rare cancers across the board, whether they're in children or in adults, cancer can occur anywhere in your body. And so while we hear a lot about the most common cancers - and the four most common are breast, lung, prostate, and colon, those four together account for about 75 or 80 % of all cancers diagnosed in the United States - it leaves this sort of piece of the pie of all the other cancer types sort of all mushed together, right? For us to be able to make impact for the rare cancers or even the rarest subtypes of common cancers, we need to be working together. It's another reason why the data sharing can have such great impact.

So another big component of CBIT is the Office of the Chief Information Officer that's led by Jeff Schilling. And I believe he's been a guest on the podcast previously. Yeah.

Oliver Bogler:

That's right, at the beginning of the year.

Jill Barnholtz-Sloan: 

And they do everything IT. So they are the ones that are responsible for everyone's laptops, for the phones, for how all the technology works in the conference rooms at the NCI. But in addition, they do a lot of stuff, scientific computing, have software engineers, they oversee the platform that's used in the intramural program for data management and sharing. So they do also really a large gamut of things.

And then there's the Informatics and Data Science program. That was the role that I was brought into the NCI three years ago to oversee as the Associate Director for Informatics and Data Science. And the informatics and data science does everything data that you can possibly think of, right? So how do we define data elements? How do we harmonize data elements? How do we organize them to make them more interoperable and useful? How do we put them into a data sharing platform and make them useful for people, and have all the right infrastructure, the right tools, the right interactions for people? And then we also do analytics and develop novel algorithms. We also have a team that works on real world data and how do we use real world data in conjunction with other data sets that we have already at the NCI.

Oliver Bogler:

Define for us what real world data means in this context.

Jill Barnholtz-Sloan:

So a lot of the real world data that we would be using would come from two major sources, right? So one would be electronic health record data from different health systems, and the other would be from claims data, like Center for Medicaid and Medicare services type data.

Oliver Bogler:

So these are clinical data, but not coming directly out of a structured clinical trial, not collected in the context of a trial.

Jill Barnholtz-Sloan:

Right.

Oliver Bogler:

They're being collected for other purposes, but can then be used for research.

Jill Barnholtz-Sloan:

Right, which is why they're called real world, right? Because they're collected for regular healthcare purposes, not for the purpose of research. Whereas a lot of the other data assets that we help make available were collected for research purposes. A lot of those research collected data have definitely been utilized to learn more about how to better diagnose certain cancer types, how to identify novel drug targets for cancer types. Even though they were collected as research data, they were used to make discoveries that had clinical impact.

Oliver Bogler:

So again, I think the theme of trying to maximize the value of data that is being collected for a variety of purposes by sharing it, organizing it, and making it accessible so that lots of people can use it to advance their research programs.

Jill Barnholtz-Sloan:

Yeah. And I think, you know, the other interesting thing is that we've started to really incorporate patient advocates in, in some of our efforts, and allow them to be, you know, users and get their feedback. And it's just, you know, very interesting and really, touching how a lot of the patient advocates really want their data shared, right? And it's very important to them and so I think for all the reasons we are discussed about data needing to be out there because it's being generated faster than a lot of people can analyze that and you know this all the different people around here are doing novel things with that that need more data in order to validate findings to make new discoveries but in addition to that it's important because we owe it to the individuals with cancer and their families.

We have one other group within CBIIT. We have an Office of Business Operations, and they sort of help oversee all of the contracts that we utilize to do and lots of other things, communications, finance, all of that, all the business operations components for all of CBIIT.

Oliver Bogler:

So what are some of the biggest challenges and opportunities that you see in this domain of data science and informatics? And particularly from the vantage point of being the leader of these efforts at NCI, what are the NCI's opportunities to make progress?

Jill Barnholtz-Sloan:

I'd say a couple things come to mind. the first is workforce development. I think it's because data is being generated so fast and furious, it's a really great, whether it's coming from the computer science side or it's coming from the statistical side or the epidemiology side or now there are programs of data science, right? There never used to be, but I think there's a lot of job security there, right? Because there's so much data being generated as we go through time, there's going to be even more and more jobs available. And so what part should we be playing as the NCI in helping to develop, you know, the workforce and in data science, data sharing and policy and, and IT, I think there are a lot of opportunities there.

I think the other really big opportunity is to help with a culture change of helping investigators, whether they're intramural at the NCI or extramural, right? Any investigator in the cancer space to make sure that they are thinking of themselves not just as a data generator, right? So in other words, I have my grants, I have my project, I generate data, right? I use those data for discoveries and publications, but also thinking of themselves as a data consumer.

And the big distinction in my mind there is if you also think of yourself as a data consumer, then you're thinking a little bit more at the front end before you start to generate data about how do I want to organize the data? What other data sets might I want to analyze my data set with? And that would help, that would maybe change how you're organizing the data or how you're defining different data elements, right? So I want to help the whole community understand that they're not just generating data, but they're consuming data and those two are really important to go together.

Oliver Bogler:

Interesting. So your own research has focused on brain tumor epidemiology, and I think you have a research program in the Division of Cancer Epidemiology and Genetics at NCI. Tell us about that program.

Jill Barnholtz-Sloan:

Yeah, so I started, so my PhD is in biostatistics. And when I was in grad school, the human genome was about to be published. I don't have to tell anybody, you can look it up what year that was so that now you know all day, that's okay. So it was late 90s, right, close to 2000. And my academic advisor in my PhD program said, you better learn this stuff. This is the next stuff that's coming. And I said, geez, you know, I'm getting a PhD in biostatistics. Like, isn't that enough, you know? And he said, no, it's not enough. So he pushed me to take human genetics and population genetics and molecular biology so that I would be at least familiar with the vocabulary, right, and have that opportunity to engage with investigators in that space and be able to talk with them.

Then he said, well, you're going to have to make some decisions about what you want to do with your career, you know, and, a lot of the other folks that graduated with me, they chose, I would say a very standard route for a biostatistician at that time, which is, know, to do clinical trials. Right. And I knew that that was something that is obviously critically important, but not something I was really that interested in having as a career as a clinical trial biostatistician, even though I did that, of course, with people when I was in academia, helped them design their clinical trials and so forth. So I got a research fellowship at MD Anderson. That's when we met, Oliver, right?

Oliver Bogler:

Yep.

Jill Barnholtz-Sloan:

And I worked for a professor who's a brain tumor epidemiologist. And so, know, really started, I would say very, very early on in my career doing brain tumor epidemiology. At that time, we were still doing etiology studies, right? We were using a case control design to try to identify risk factors for brain tumors. I think the brain tumor epidemiology space has changed directions a little bit because I think many of us feel that we've exhausted so many different potential exposures, know, looking at those different exposures and it's sort of come up empty handed, unfortunately, and really have not been able to identify a risk factor that explains a large number of brain tumors.

And so a lot of us have switched to looking at molecular characterization of the tumors and what's the relationship between the molecular features of the tumors with response to treatment, other types of clinical outcomes. And so a lot of the work that we focus on with my small but mighty team in the division of cancer epidemiology and genetics is looking at sex differences in brain tumors, and also just we've gotten together a group from around the NCI also with the Office of Women's Research and Health, looking at sex differences in cancer globally. So we just hosted a series of virtual workshops focused on sex and gender differences in cancer. And those recordings are available for people to watch if they're interested. Two of them were focused on basic science. One was focused on clinical science and one was focused on population science. And our next steps with that are to write some white papers and try to get some papers published.

Oliver Bogler:

So we'll put some links in the show notes to those recordings that you mentioned.

Jill Barnholtz-Sloan:

That would be great.

Oliver Bogler:

Of course. But I'm curious, what was the finding for a for gender and sex in in glioblastoma?

Jill Barnholtz-Sloan:

Ah. Well, brain tumors are really complicated. There are malignant brain tumors and then there are non-malignant brain tumors. The hallmark molecular features, the types of treatment they got, the outcomes, totally different. And then within the non-malignant type and the malignant type, there are many, many, many, many, many subtypes.  And some of those subtypes are now defined by hallmark molecular feature.

So in other words, you have to have this molecular feature in the tumor to be able to call it this type of brain tumor. Do those hallmark molecular features define a treatment choice? But we're getting there, right? Which is great. And a lot of other cancers, as you know, if you have a hallmark feature, then you get a targeted therapy that hits that whatever mutation or copy number change, et cetera, that you might have.

So glioblastoma is the most common type of malignant brain tumor. It’s very common in adults, very uncommon in children. So brain tumors, if you just look at the malignant piece of brain tumors, there's a bimodal age distribution. So they are the most common type of solid tumor in children. Of course, the most common type of cancer in children is leukemia. But brain tumor is the most common type of solid tumor in children. So there's an incidence peak in childhood between sort of zero and five years old, so newborn to five years old. Then it goes way down and then it starts to come back up again, 60s, 70s and 80s for the malignant brain tumor types. The non -malignant brain tumor types really don't occur in children very commonly. They occur much more commonly in adults.

It depends on which type of brain tumor you're talking about, whether or not there's a sex difference. Globally for all cancers, males get cancer more often. So this is excluding any cancer that would be specific to a biological male or female, right? So it's not including female breast or male prostate cancer. So it's all the other types of cancers - that the non-reproductive cancers. Males get cancer more commonly, but their survival overall is worse than females. And we really don't fully understand why that is. So for glioblastoma, it follows the same pattern. Glioblastomas are more common in males, and males have worse survival.

Oliver Bogler:

Interesting. guess lots more work to be done to tease out the reasons for that.

Jill Barnholtz-Sloan:

Yeah, and the gender differences are very hard because gender is not typically systematically captured. I think some of the different health systems are trying to capture that information now. So we really know even less about gender differences. When I say sex, mean biological sex, right? Male versus female, what sex were you born? Yeah.

Oliver Bogler:

Assigned at birth, got it, yeah, understood. Interesting. So Jill, you've been a sought after expert in informatics and data science, and I wonder what your thoughts are. What are some of the essential skills for researchers who want to leverage the data that you talked about? They're everywhere. What do you need to be an effective researcher that is able to access and make use of those data in the fight against cancer?

Jill Barnholtz-Sloan:

Great question. And it's something that we grapple with, with the data sets that we have available and make available, right? I would say the bulk of our current users of our NCI, you know, sort of data assets, right, are folks who are computationally or quantitatively inclined, right? So they're gonna be folks who would define themselves as data scientists, statisticians, computer scientists, computational biologists, quantitative biologists, even there's lots of different terms that people use, bioinformaticians.

But we are working very hard to lower the barriers to being able to interact with the data. We're rolling out a new data submission portal for the Cancer Research Data Commons, which is where a lot of our NCI data assets sit and are made available. In addition, we're gonna have tools that are on top of the data that are gonna be as easy to use as your web browser would be or a Google search.

So we're gonna have the ability to interface with the data in a very easy to use way, but that's coming down the line. I think in order for us to truly make all of the data assets available to the full cancer research community, which includes researchers, clinicians, individuals with cancer and their families, trainees at all levels, it's our responsibility to make those data as easy to interact with as possible. So we're working on that.

But I guess I would say is, if you're quantitatively inclined, I would highly recommend that you look into data science or computational biology or bioinformatics and ask questions and learn more about it, right? Reach out to some professors, you know, if you're still in school and talk to them about it. I'm happy also to talk to anybody if they want to reach out to me.

Oliver Bogler:

We'll put your contact info in the show notes as well.

Jill Barnholtz-Sloan:

Okay, great. And I know, Oliver, you talk with folks all the time, you know, about career paths and choices. I would also say it's interesting because it seems to me that a lot of basic science labs and translational science labs and even clinical labs are starting to cross train their trainees with at least helping to enable them with some more quantitative knowledge and experience. So I think to be coming out of a lab and you know how to do the basic science you know, what I would call wet lab experiments, you understand all the nuances about the experiments and all the potential biases that come into play. So you have a really deep understanding when the data come out of how they regenerated and what all the nuances are. If you can then marry that with some quantitative skills, I just think it's just a great, great thing to have both of those together.

Oliver Bogler:

Yeah, so just to dig a little deeper. right now, I mean, I think the perception is that in order to really meaningfully explore data, set this large data sets that you've been talking about, you really have to have not only some quantitative interest or capability, you also have to have some facility with computational tools like Python or R and obviously statistical methods and so on. And I agree, those things are being more integrated into general cancer biology training programs and other venues, but can I dream of a future where, say an AI sits between me and the data and all I need to know is a question, and it will select the tools and do the heavy lifting for me?

Jill Barnholtz-Sloan:

Yeah, I think that there are a lot of things that are moving in that direction, right? Where just the way that you would ask Google a question and it would give you back answers, we are hoping that we're gonna have a similar kind of interface. I don't know if it would be a Google, you know, I'm saying Google-like, but what I mean by that is it's just so easy to sort of type in a question and you get an answer back. Sometimes you don't get the answer you want, right? So they're not perfect.

I think all of the AI and machine learning algorithms are super interesting, right? But just standard statistics approaches, statistical approaches, simple things like a t-test or a non-parametric t-test or, you know, analysis of variance or a regression analysis, they can be extremely powerful and we understand what the assumptions are of those tests. We know how to test for those assumptions in our data and we understand what the interpretation is of those results.

So I understand how you can kind of see stars in your eyes when people are talking about AI and ML, but I think it's important to understand that there's still a little bit of black boxiness with those algorithms. And it can be very hard to understand what are the assumptions of those algorithms? Do your data fit those assumptions? And then what is your interpretation of what comes out can be very challenging. So what I would say is don't worry about learning the AI and ML stuff. Make sure that you learn the basic statistics, I think is really important. And you can use Excel to do the basic statistics. You don't have to know R or Python.

Oliver Bogler:

So there's really no substitution for understanding the fundamentals, even in the future when you might use tools like AI, because you have to be able to have confidence in what you're seeing.

Jill Barnholtz-Sloan:

Absolutely, I you have to understand like, how do I look at the distribution of my data? What does a median mean? What does a mean mean? What's the difference? If I have two sets of data, how do I compare them? What sorts of tasks could I use? You don't need a fancy task to do that, right?

And I tell everybody on my team over and over and over again, like they're just, sick of hearing it is before we do any kind of statistical tests or any modeling, we have to look at the frequencies, right? We have to understand the data that we have generated, the data that we have, because it's the only way then to figure out do we have certain pieces that are missing that we don't know why they're missing, right? Or, are there some weird things when we look at the distribution of the different variables that may clue us into something happened with the data that was out of our control? It's just the easiest way.

So it's like, I tell everybody, we start with the table one. Table one is your frequency table in your paper. Start with the table one. Don't run any statistical tests. If you show me a table that has t-test p -values in it, I'm gonna tell you, I'm gonna ask, where's the table one?

If it's too daunting for somebody who's a basic scientist, a clinical scientist, a translational scientist to say, my gosh, I have to learn R or Python or learn all these AI and ML procedures, while I think it's great if you want to learn that, absolutely, I would never discourage anyone. Make sure you understand sort of basic statistics and the different sorts of tests and regression approaches that would be used in medical sciences, I think. It's a pretty small list of things that we use over and over again.

Oliver Bogler:

Good advice. Thank you so much. We're going to take a short break. And when we come back, we will talk about Jill's career path.

[music]

Have you ever thought about harnessing the power of data science in your cancer research, but aren't sure where to start? The good news is that you can take a step toward unlocking the potential of data science with the National Cancer Institute's new Cancer Data Science Training, your gateway to free online training and resources to help you learn how to work with big data and deal with big data problems in cancer research.

Go to datascience.cancer.gov/training and there you can take the Data Science 101 video course to learn the basics about data science skills such as coding, machine learning, biostatistics, and more. Or explore the Cancer Data Science project lifecycle where you'll learn how to start your own projects and see where you fit in as a cancer researcher.

NCI experts wrote six basic guides with tips you can apply to your own research. Or explore the training guide library, your easy access toolbox of resources for cancer data methods, tools, and technologies.

Use it to get quick answers on identifying and gathering data, predictive modeling, data visualization, and more. Data science is accelerating discovery in cancer research. Head to datascience.cancer.gov/training and be a part of it today.

[music ends]

Oliver Bogler:

All right, we're back. Jill, you've had such a multifaceted career in biostatistics, data science, and cancer epidemiology. Could you share with us what initially sparked your interest in science and then what led you into these fields? And I know you've shared a little bit about your growing interest in brain tumors, but when was the first time you thought, I'm going to be a scientist?

Jill Barnholtz-Sloan:

Yes, I didn't think that at all. I thought I wanted to be an architect. And I'm very grateful to my parents. They saved enough money to send me and said, okay, just pick where you want to go. Right. And so I was looking at undergraduate universities based on their architecture program, right. The ranking of their architecture program.

And I also, grew up in St. Louis, Missouri, you know, very Midwest kind of St. Louis is not a big place. And I really wanted something different. and I didn't want any winter. That was another, qualification. So I mostly was looking at schools in the South and then looking at their rankings for architecture programs and university of Florida at Gainesville has an exceptional architecture program. And it was in the South, right? So, know, myself, like anybody else, applied to multiple different schools. And so I got into Florida and I said, I'm going, you know. So I get there and I start my first semester. And of course you start immediately in, these architecture design classes, right? And I soon realized that I was in way over my head.

So we would go, we would go to class. And the professor would say, okay, we're walking to this location on campus. And we would get to said location and it would be an open space in between two buildings. And he would say, design a multipurpose living apartment building for families. Go, you have 24 hours to bring me back, you know, a … you had to build a model, right? And I just thought, oh no. I I really thought it was, I've always been a quantitatively inclined person. It's just sort of how my brain works. And I always thought architecture was more mathematical. It's really design, you know? I mean, there's mathematics, of course, involved in it, but it's really design. And so I thought, this is just, this is not gonna work, right?

So I made it through the semester and then I was like, I've got to switch majors. So I switched to civil engineering. University of Florida also has a huge engineering program. And it seems like the right thing to switch to because all of the prereqs were the same. all of the classes I had taken that first semester just immediately were the same ones I needed for an engineering degree.

That first summer I started working for a small civil engineering firm in Gainesville. I stayed in Gainesville. I took a class. I worked for this engineering firm. They had just started using computer aided design, so AutoCAD. And so I was an AutoCAD person. I was taking their designs that they had written by hand for highways and roadways and putting it into the computer. And I soon came to realize that a lot of the engineers who were great people and so supportive, they had been doing the same thing for 15 or 20 years, like literally the same thing. And I thought, uh oh this is, I don't think this is for me, you know, at all. So then I thought, well, what else could I do? So now I'm a year into school going out of state, right? And my parents were like, look you gotta pick a major and we gotta be on a track here. You know, you can't be there forever. Which now as a parent, I understand, but you know, at the time I thought they were being unfair.

So I switched to mathematics and that's when I ended up getting my bachelor's degree in. And my path to statistics was literally as random as well, I took two semesters of mathematical statistics that were required for my mathematics degree. I liked it, so I'll go get a master's in statistics. So that's what I did. And then when I was doing my master's degree, I realized that there were two ways you could go, right? Applied statistics or theoretical statistics. Theoretical statistics for sure was not for me. And then I started to explore the differentareas of applied statistics which really are business statistics which is a lot of forecasting and sort of econometrics, what I think would be really interesting. Engineering statistics which has a lot of overlap with industrial engineering so it's a lot of quality control and those types of things. Quality assurance type analysis and then medical statistics or biostatistics. So then I applied to the biostatistics PhD programs and that's how I landed at University of Texas School of Public Health in Houston. That was before they had a lot of all the other schools of public health around Texas which they have in lots of other cities. And actually, so the department at the time was called biometry which is a very old-fashioned term for biostatistics. They changed it to biostatistics and now it's called I think data science and biostatistics, and they are celebrating the 55th anniversary of the department. And I'll be going in November for a celebration, which is pretty exciting.

Oliver Bogler:

Nice. Nice.

Jill Barnholtz-Sloan:

So yeah, so that's how I ended up there. And then I think I really had to make the choice about, know, do I want to go work for like a big pharmaceutical company as a PhD level statistician? That's what a lot of my colleagues ended up doing or do I want to be, you know, doing research in academia? So, you know, grateful to the research opportunity I had an MD Anderson, which really allowed me to see what that was like. And then obviously I chose the academic path and was at multiple different universities, mostly following my husband where he, where he went. But I think I made lemonade out of lemons, I guess you could say. I took advantage of every opportunity I was given. And then about three years ago, this opportunity at the NCI came up and I thought, this is a pretty cool opportunity to have, you know, to be able to be a part of something bigger than myself. mean, NCI is the mothership of cancer, right? 

Oliver Bogler:  

Right, right. And I heard you in a retreat that we recently had relate the story that some of your academic colleagues, when you announced that you were coming to the NCI, were, why would you take a government job?

Jill Barnholtz-Sloan:

Oh yeah, they were like, you're going to work for the government, right? Like big brother or something. I said, look, you know, I have a very close friend who does a lot of career counseling for folks at the, I was at Case Western Reserve University School of Medicine. And I had a lot of conversations with her where she was forcing me to think through the really hard questions, right? About where do you want to go? What do you wanna do? Like, there was no reason for me to leave Case Western. Like, I could have just continued doing what I was doing. But I had, I came to the realization in all of these conversations with her and of course talking with my husband that there were other things that I wanted to do with my career. And when this opportunity came up with the NCI, it checked some boxes for me of some things I wanted to do. So I was lucky that they chose me.

Oliver Bogler:

So what were those boxes?

Jill Barnholtz-Sloan:

Well, I was looking for a new leadership opportunity, a different type of leadership opportunity. I was looking for an opportunity to really push the boundaries of what I knew I could do in terms of management, strategic thinking, and kind of taking something that the program, the Informatics and Data Science program, all the pieces were there in CBIIT, but they weren't in this single program. They made the single program and then hired in me as the associate director to oversee the program.

So there were a lot of interesting things to do, a lot of strategy about staffing and budget planning and those types of things, which as a single PI, you know, in the extramural space, you oversee kind of the money and the staff that you bring in based on the grants that you're bringing in. But you don't get this type of management opportunity. And so I was really interested in that. I was also just honestly really starstruck about the idea of working at the National Cancer Institute and just being at a place, since I've been doing cancer since I was a graduate student and nothing else, being at a place where everybody was focusing on cancer seemed really cool to me. And then I was lucky enough to be able to negotiate to still have an intramural research appointment so I could still do research.

Oliver Bogler:

So closing out this part, what advice would you have for our listeners, for people early on in their careers? Maybe they are mathematically inclined and interested in cancer research. What advice would you give them? How should they shape their path to come into the area where you have worked for so long?

Jill Barnholtz-Sloan:

Gosh, that's a great question. Well, I think the first is like, honest with yourself about what you think you're good at, right? Where you think maybe some gaps are that you would like to fill, you know, with additional training. And then how you might want your lifestyle to be, right? When you're working, because if you're in government, if you're in academia, if you're working in industry, the lifestyle and what I mean by that is just the stressors that you have, you know, can be very, very different. And, you know, how do you find out what those are, you talk to people and you ask questions and you gather your information and then based on the information that you've gathered, you make the best decision for yourself.

You know, give yourself some grace, right? Sometimes you gather all this information and you make a decision and you get into it and you think, what have I done? So be kind to yourself and realize that you can pivot and it's okay. And sometimes pivoting is really scary, right? When I pivoted to work for the government, that was a big pivot for me. That was a big pivot for me after being in academia for over 20 years and sort of understanding how that game worked, you know?

Oliver Bogler:

You're three years in, any regrets?

Jill Barnholtz-Sloan:

No, not at all, not at all. I miss my colleagues in Cleveland. We were in Cleveland for over 16 years and I had really just this amazing network of colleagues from all different disciplines and all the different institutions in Cleveland. And I do miss that. It's hard to go from being somewhere for a long time and really having this big network of people to starting somewhere new and you have no network, right? So, but no, I do not have any, any regrets.

Oliver Bogler:

I can empathize having come in from the outside about five years ago. I still remember that.

Jill Barnholtz-Sloan:

I hear you're allowed to say you're new in any position for up to five years. So I still have two more years to go.

[music]

Oliver Bogler:

Now it's time for a segment we call Your Turn because it is a chance for our listeners to send in a recommendation they would like to share. If you're listening, then you're invited to take your turn. Send us a tip for a book, a video, a podcast, a talk, anything that you found inspirational or amusing or interesting. You can send these to us at NCIICC@nih.gov. Record a voice memo and send it along and we may just play it in an upcoming episode. Now I'd like to invite our guest to take her turn, Jill.

Jill Barnholtz-Sloan:

Thanks, Oliver, I appreciate it. So I was thinking about this, because you had asked me to think about something. And for my turn, I would love to say set boundaries so that you can keep for yourself some kind of balance in your life. So the boundary that I set, you know, obviously I switched to a new job by coming into the government. Whenever you switch, it makes it easier to set a new boundary versus when you've been somewhere for a long time and you haven't set boundaries at the beginning, it becomes more difficult over time to set boundaries. So the boundary that I set is every day between 12 and one, I really try my best not to have any meetings during that time. And I get out and walk the dog, and have something to eat or put the laundry away or whatever, but it really gives me a nice mental breather in the middle of the day.

And I find that on the days where I do have a meeting, know, set during that time that I have to be on, I can sense within myself that I missed something, you know, what did I miss? And then I realized it was that mental break. That's my, the thing I would like to really stress to people that, you know, you have to be an advocate for yourself and you shouldn't ever, ever, ever feel badly about doing that.

Oliver Bogler:

I think that's a great, great piece of advice. Thank you. Thank you for sharing that. I'd like to make a recommendation as well. Yes, my obsession with AI continues for people who listen regularly. So I'm recommending today an online learning about neural networks, the technology that is central to so much that is going on inside AI. With some neurobiology in my background, I felt I had a reasonable grasp of some of the likely characteristics of these computer neural networks.

But when a colleague recommended the website 3Blue1Brown, and specifically their neural network course, which is quite informal, I began to learn how these things really work. With engaging visuals created with an animation engine that Grant Sanders, the person behind 3Blue1Brown… he explains what is going on under the hood with machine learning. It's only a tiny part of what is offered on his YouTube channel, also called 3Blue1Brown, which has over 6 million subscribers. There's really a ton of stuff in here, but I've been zeroing on this one learning path about neural networks. And it's really interesting. So if you're as fascinated with AI as I am, but also want to learn a little bit more about what's going on under the hood, then I can recommend this site. And of course, we'll put a link in the show notes.

Well, Jill, thank you so much for spending time with us, sharing your work and your career path and your advice with us. Thank you.

Jill Barnholtz-Sloan:

Thank you so much for having me.

Oliver Bogler:

That’s all we have time for on today’s episode of Inside Cancer Careers! Thank you for joining us and thank you to our guests.

We want to hear from you – your stories, your ideas and your feedback are welcome. And you are invited to take your turn and make a recommendation to share with our listeners. You can reach us at NCIICC@nih.gov.

Inside Cancer Careers is a collaboration between NCI’s Office of Communications and Public Liaison and the Center for Cancer Training. It is produced by Angela Jones and Astrid Masfar.

Join us every first and third Thursday of the month wherever you listen – subscribe so you won’t miss an episode.

If you have questions about cancer or comments about this podcast, you can email us at NCIinfo@nih.gov or call us at 800-422-6237. And please be sure to mention Inside Cancer Careers in your query.

We are a production of the U.S. Department of Health and Human Services, National Institutes of Health, National Cancer Institute. Thanks for listening.