Asst. Prof. Ercüment Çiçek of the Department of Computer Engineering holds BS (2007) and MS (2009) degrees in computer science and engineering from Sabancı University, as well as a PhD in computer science from Case Western Reserve University (2013). While pursuing his doctorate, he spent time in 2012 at Cold Spring Harbor Laboratory to work on gene discovery algorithms for autism spectrum disorder. After graduation, he worked as a Lane Fellow in computational biology at Carnegie Mellon University (CMU) until 2015, when he joined Bilkent University; he remains an adjunct faculty member in the Department of Computational Biology at CMU. Dr. Çiçek’s research focuses primarily on the design of machine learning algorithms for analysis of large-scale biological data. He is the recipient of a Simons Foundation Autism Research Initiative Explorer Award, a TÜBİTAK Career Award and a Turkish Academy of Sciences Outstanding Young Scientist Award.
How did you end up here, at Bilkent?
While doing my postdoc at CMU in 2015, I came to Turkey to visit my family. I gave one talk at Bilkent, and they were interested in my research and said, “Why don’t you come?” I thought Bilkent was a nice place, so I said yes. It’s like that; I just ended up here.
What do you like about being at Bilkent?
I think it’s one of the best universities in Turkey, especially for my profession, computer science. We have very good students, at both the graduate and undergraduate levels. It’s very dynamic. There are many good professors as well. The campus is really nice too, with the woods. It’s very different from the rest of Ankara. I really like it; I like being here.
Why did you become a scientist?
I had an internship and decided I didn’t like the industrial sector. I thought the project I was working on was really boring. I like the independence aspect of academia better. I have the freedom to do research on whatever I want; I don’t have a boss, in that sense, who tells me what to do. I don’t like waking up at 8 a.m. and going home at 5 p.m.; I like flexible hours.
Why computer science?
I was always interested in computers. I went to Sabanci University for my undergraduate studies, and there, we were able to select our field at the end of the second year. So, I liked computers, but I didn’t know for sure; I kind of inspected mechatronics engineering and industrial engineering as well. But I finally decided on computer science because I like to solve problems in my mind. I don’t like to build things with my hands, like a mechatronics engineer would build a robot. I like to abstract something in my mind and then tell the computer to do it. So, that’s how I ended up being a computer scientist.
Could you explain your current research for our readers?
I’m a computer scientist, but I’m working on problems related to biology and genetics. There are many things that I work on, but currently, I’m mainly working on the genetics of neurodevelopmental disorders. These are complex diseases. In a nutshell, the human genome has 20,000 genes, and we suspect, for instance, that for autism (my main interest), there are 1,000 genes, and combinations of those, that are responsible. If you think about the combinations of 1,000 genes, and you don’t know what those genes are, it is quite heterogeneous and difficult to know what genes are responsible. So, currently we know 60 genes; we need to get to 1,000 genes. Normally, we need a large number of patients to sequence, which is very costly and time-consuming. So what I’m trying to do is to build computational models – methods that would predict those genes without the need to sequence patients. We’re trying to do it ahead of time, and then the main goal is to be able to use those genes to be able to tell something about the outcome of the disease in patients. Basically, it’s building machine-learning algorithms for problems in genetics.
How did you become interested in autism?
While working on my PhD, I did an internship at the Cold Spring Harbor Laboratory, founded back in the day by James Watson (one of the guys who discovered the double helix structure of DNA). This was in 2012. My internship was with a professor there who was working on algorithms for this problem. We developed the framework there, and I really liked the topic. It’s very interesting, very complex and very suitable for building methods. Also, it’s nice to have a sense of helping families affected by autism – maybe not now, but at least in the future. And then I did my postdoc in a competitor group, which was working on the same thing. They liked me because I had experience in the field, and so I ended up at CMU. Since I liked the topic, I continued working on it after I founded my lab.
What’s the coolest thing about your work?
Being a professor is nice because I’m always around young people who are full of energy. It’s not just a workplace with a small number of people; it always changes, as students come and go. And you teach them something small and forget about it, but unlike in any other profession, they remember it and then years later they tell you that it was really nice and you really helped them.
Plus, I don’t have to wake up really early. That’s another thing I like.
Moreover, my research has the potential to help people and have an impact on real life. I really like this. It’s not just theoretical. It has the theory, but it’s quite applied, and I feel like I’m doing something that’s not going to waste. Even though everyone publishes papers, having that sense of helping someone, or the hope of helping them in the future, is also very nice.
Can you share a defining moment in your work as a scientist?
During my master’s, I worked on data privacy. I thought we were doing really nice science, but people don’t use it that much. For my PhD, I changed my field, switching to bioinformatics.
I really liked the things that my future advisor was working on. It was a big decision, to be honest. Then in the first semester of my PhD, they made me take biochemistry with the pre-med students, with nothing beyond an introductory background in biology. It was tough, but I’m glad I did it. That was the defining moment.
When and where do you do your best thinking?
While walking. Since my undergraduate years, whenever I’ve had an algorithmic problem, I’ve always left everything – no earphones, no nothing – and just walked around and thought about it. I always get rid of the complications and have a clear mind about the problem. Sometimes I solve it, and sometimes I don’t, because not everything is solvable. But I always go for a walk.
What problems do you hope scientists will have solved by the end of this century?
A lot. A lot. If you think about machine learning, it’s like a 30-, 40-, 50-year-old science. Think about the things it has solved up to now. I always say this during the professional seminars we have here for high school students. Five years ago, I was saying that computers are really good at things like counting the occurrence of a word in a book, but not good at recognizing faces. On the other hand, as humans, we cannot count as fast, but we’re very good at visual recognition; I can recognize faces even though I’m really bad at it compared to some other people. But now, five years later, computers are really good at recognizing faces as well. You know, Facebook can tag you in pictures, and we have face IDs in our phones. So, a lot has changed even in five years.
In 10 to 20 years, easily, I think we’re going to see autonomous cars on the streets. In fact, I think humans will be banned from driving.
And then health care is also going to change a lot. The advances are going to help people a lot. Currently, there are papers that say, “Algorithms perform better than doctors; they diagnose better.” There are catches to that, I think. But they’re going to go hand-in-hand; you’re going to have a doctor and an algorithmic doctor. The doctors are going to consult and then decide.
Really, a lot is going to change soon; I’m not thinking about, say, the year 2100. These things are going to happen quite a bit earlier.
What’s the most common misconception about your work?
People think that as a computer scientist, you have to code all the time. I do code. But as a professor now, I don’t have to code that much. The graduate students do that. I try to solve algorithmic problems. That’s my main work.
Also, people always think I have to know every aspect of a computer: the operating systems, the architecture, which main board to buy. I usually don’t have any idea. Or they ask me, “What’s the best cell phone to buy?” I don’t know much about cell phones. I mean, I have one, but that’s it. That’s the greatest misconception: If you’re a computer engineer, you have to know everything about a computer. I forget who said this, but there’s a famous saying: “Computer science is no more about computers than astronomy is about telescopes.” People don’t understand that.
What are the qualities necessary to be a scientist?
You have to be hard working; that’s a given. Also, in engineering, you have to be into math and science. One thing I really liked that one of my advisors’ advisor said is, you should always persevere, it’s not always the smartest guy who achieves, but the person who, even if he/she struggles a lot, doesn’t give up. You have to keep going no matter what. You shouldn’t stop.
Any advice for students who are starting their careers?
The first thing they should know is that university is not a place where everything is taught. This is unlike high school, where you have a curriculum and they teach you everything you need to know. Here, students are learning to learn, and they should get their hands dirty with everything. They should not wait for a course to be taught. So we give them the fundamentals here, but they should get into everything they can. And if they want to go into academia, I suggest that they get involved in research projects with professors. I have too many right now, but usually we need help with coding something or other, because the graduate students are also very busy. I try to get undergrads to help in some way, even if it’s just to see the problems we work on. They get something on their CVs, and reference letters from the professors. This is important, because once you graduate, all the CVs look alike; the only thing that can distinguish you is your GPA. But having that extracurricular stuff, showing that you cared, that you were curious, and then you were successful, is very nice.