menu ☰
menu ˟

“Show them they were wrong to think of you under the header ‘woman’ before they thought of you as an expert”

20 Jun 2017

In honour of International Women in Engineering Day on 23rd June, Chloe-Agathe Azencott is June’s featured Faculty Member of the Month. She has been a research scientist at the Centre for Computational Biology of MINES ParisTech and Institut Curie, Paris, France since 2013, focusing on the development of methods for efficient multi-locus biomarker discovery. Problem solving and seeking new knowledge drew her into engineering.

Chloe-Agathe received an MSc in engineering and an MSc in computer science and applied maths from Telecom Bretagne, Brest, France. She then went on to obtain her PhD in computer science in 2010 at the University of California, Irvine, developing machine learning methods for drug discovery. She then spent three years as a research scientist at the Max Planck Institutes for Developmental Biology and Intelligent Systems in Tübingen, Germany. 

In particular, Chloe-Agathe is interested in the incorporation of additional (structured) information, for example as biological networks; in multi-task approaches, where one addresses multiple related problems simultaneously; and in the development of fast but accurate techniques to address these issues.

Can you tell us a little bit about your work?

I’m a computer scientist by training, and my research is on the application and development of machine learning methods for data from the life sciences. I guess the cool way of saying this would be “artificial intelligence for precision medicine”.

The breakthrough came a few weeks into my MSc internship in a research lab. That was the most fun I had had in my years doing science, and I wanted to keep doing that!

My goal is to make sense of data with a small number of samples and a large number of variables. These variables can be clinical variables, such as age, cholesterol levels or smoking history, or genetic variables, which can include gene expression, mutations, or epigenetic markers. How can we find out which of these variables plays a role in a particular biological process or pathology?

My work has numerous applications, particularly in developing treatments that are adapted to the genetic specificities of patients, by contrast with a classical one-size-fits-all approach. One of the specific areas I work on is building methods to analyse data from Genome-Wide Association Studies. In other words, I want to identify, among long lists of common mutations of a single base-pair of the genome, the ones that explain a specific observed trait, such as cancer susceptibility or how well a patient responds to a specific treatment.

Statistically, this is tricky because we have many more such mutations than patients, and this makes the data very different from the majority of what people call “big data” nowadays. There are so many possibilities, and you don’t have enough patients to rule out enough of these possibilities. The way I approach this is to incorporate additional information, such as gene-gene interaction networks or gene pathways, so that the hypotheses we generate from the data are consistent with other things we know about the biology of the problem.

 

What triggered your interest in research?

I don’t think I had a key “a-ha” moment. I remember in high-school, I liked the idea of doing research, but I had no clear idea of what it was about. To me, research represented what I enjoyed about school, learning new things and using them to solve problems. Then I ended up in an engineering school without knowing exactly what an engineer was either, but the idea of research stuck with me.

I had the possibility to work towards a research-oriented MSc in parallel with my engineering program, and it seemed like a cool way to learn more things, so I decided to do that. The breakthrough came a few weeks into my MSc internship in a research lab. That was the most fun I had had in my years doing science, and I wanted to keep doing that!

 

What would be one of the biggest challenges that you have overcome to get to where you are today?

I want to identify, mutations that explain a specific observed trait, such as cancer susceptibility or how well a patient responds to a specific treatment.

Impostor syndrome. On some days, it’s still a challenge. My daily life is made of applied mathematics, statistics, genetics, and a touch of organic chemistry. My education was in computer science. Someone will eventually realize that I have no idea what I’m talking about.

 

What was your last recommendation on F1000Prime and why did you pick it?

My last recommendation was “Improving polygenic risk prediction from summary statistics by an empirical Bayes approach” by Hon-Cheong So and Pak C. Sham in Scientific Reports. I usually focus on identifying a small number of mutations that can be related to disease risk. This paper is more interested in making accurate predictions, and not so much in interpretability, showing it’s fine to use the entire genome if we can make accurate predictions.

I think there are opportunities for bridging the two types of approaches. An important aspect of their work is that they don’t work with the raw data, but with summary statistics, that is to say, the scores given (potentially by other researchers) to the variants. Because there is no privacy concerns attached to this type of data, the summary statistics are much easier to share.

 

What would you say is the best piece of career advice that you received that you would like to pass on to early career researchers, and to encourage young women into software engineering?

Go ahead, make the most of it and show them they were wrong to think of you under the label ‘woman’ before they thought of you as an expert.

Build, create and maintain a support network. Family, friends, classmates, labmates, fellow researchers you’ve met at conferences, people from the Internet: a variety of awesome people who think you’re awesome, and people with whom you can identify. It’s about mentors, but also role models, friends who’ll relate to your problems, and people who find it ‘cool’ that you’re a scientist. Meeting other women, or minority, scientists is key, and can be greatly facilitated by the Internet, especially with Twitter or blog networks, but also women-in-science type of events, whether meetups or special sessions at conferences in your field.

I have to add, even if you think or know you’ve been invited to an event, either to give a talk, chair a conference session, organize a workshop, sit on a thesis committee, “just” because you’re a woman and the organizers want balance, don’t let it phase you. You’ve been given this opportunity. Go ahead, make the most of it and show them they were wrong to think of you under the header “woman” before they thought of you as an expert.

 

Click here to view the full article which appeared in F1000 Research