Preamble about Interdisciplinary Research
We are not students of some subject matter, but students of problems. And problems may cut right across the borders of any subject matter or discipline. — Karl Popper
The piece of work you are reading should harvest the fruit of interdisciplinary research conceived in an interdisciplinary environment of Center for Interdisciplinary Research in Paris (CRI) in École doctorale Frontières du Vivant (FdV) and Institut Curie in groups Computational Systems Biology of Cancer and Integrative Biology of Human Dendritic Cells and T-cells. CRI’s main mission can be formulated as follows:
to empower the students to take initiative and develop their own research projects at the crossroads of life, learning, and digital sciences. (“The CRI | Centre for Research and Interdisciplinarity” 2018)
Interdisciplinarity has many definitions and meanings. According to the book Facilitating Interdisciplinary Research (Facilitating Interdisciplinary Research 2004)
Interdisciplinary research and education are inspired by the drive to solve complex questions and problems, whether generated by scientific curiosity or by society, and lead researchers in different disciplines to meet at the interfaces and frontiers of those disciplines and even to cross frontiers to form new disciplines.
For me, the essence of interdisciplinarity is the need to solve a complex problem, whatever expertise would be necessary to solve it. I consider that fighting cancer disease, deciphering cancer heterogeneity and interactions of the immune system are causes worth an interdisciplinary effort. This is even truer in the era of big data when the demand for quantitative tools is exponentially growing, in order to extract information and knowledge.
Though this preamble I would like praise not only the interdisciplinary research but also underline possible limitations and constraints that come with it and which could affect this thesis.
What does interdisciplinarity in science mean in XXI century?
In the ancient history, being formed and practice multiple disciplines was not anything unusual which is strongly reflected in Greek philosophy initiating the dispute about the division and hierarchical classification of knowledge. (Slavicek 2012). Figures as Aristotle and Leonardo Da Vinci that can be called homo universals served different disciplines from arts through history, natural sciences to mathematics. With time human knowledge about the word, i.e., natural sciences got bigger and bigger, to the point that it became hard to master all the disciplines. The specialization would allow to study in deep a certain subject and make possible discoveries about it. And even if, interdisciplinary efforts never stopped, for a long time they were not mainstream in scientific communities divided into academies, chairs, and specialization.
Different fields differ in term of concept, method, tools, processes, and theories (Slavicek 2012). Thanks to division into scientific disciplines a sort of order is conserved across space and time. Hierarchical classification of knowledge comes from human nature.
It can be observed that there is an increasing gap between disciplines along with specialization.
advancing specialisation leads to gaps in the level of comprehension between individual disciplines and eventually gives rise to the demand for interdisciplinarity - in order to close the gaps between disciplines.(Slavicek 2012)
It is not really clear why this gap must happen. Would it somehow reflect human nature, the strong need to divide things into discrete categories rather than to see a continuum?
Nowadays, the knowledge is accessible, and we can profit from achievements of different disciplines thanks to easy means of communication. Two different terms can be defined to describe initiatives that use the knowledge of different specialties: multidisciplinarity which is a sum of efforts of different disciplines and interdisciplinarity that allows profiting from the synergy of multiple disciplines (Fig. .). With interdisciplinary research and education come flexibility, creativity, and novelty but also limit of depth on ingested knowledge and possibilities of cross-interactions between disciplines.
Figure .: Symbolic illustration of a sum (multidisciplinarity) versus synergy (interdisciplinarity), in an interdisciplinary project sum of thee disciplines A, B, C should have more value than a simple sum of disciplines: an interdisciplinary project should have an added value compared to a multidisciplinary one.
Why are not all of the labs interdisciplinary?
Scientists tend to resist interdisciplinary inquiries into their own territory. In many instances, such parochialism is founded on the fear that intrusion from other disciplines would compete unfairly for limited financial resources and thus diminish their own opportunity for research — Hannes Alfvén
Crossing frontiers is not an easy task, and it was quite difficult in the beginnings of modern interdisciplinarity. Some examples of early interdisciplinary efforts of the 20th century are nicely described by Ledford et al. (Ledford 2015) in Nature special issue on Interdisciplinarity. It illustrates Theodore Brown in 1980s while trying to organize a new interdisciplinary research project and reorganize university space to engage an exchange between students of different faculties, and he encounters a lot of reluctance.
And then there was the stigma. “Interdisciplinary research is for people who aren’t good enough to make it in their own field,” an illustrious physicist chided (Ledford 2015).
The story seems to end up with a happy ending of 40-million US dollars grant and foundation of Beckman Institute for Advanced Science and Technology. However, recruiting an open-minded director for leading this unconventional organization was a struggle. Shortly, the structure became a model for others and met a great scientific and technological success.
Even though, since then the idea of interdisciplinary research spread around the world. Yet, not all problems were overcome.
“There’s a huge push to call your work interdisciplinary,” says David Wood, a bioengineer at the University of Minnesota in Minneapolis. “But there’s still resistance to doing actual interdisciplinary science”.
First, the institutions, universities where research is performed should equip scientist with a passport to other disciplines, facilitate exchange, funding the interdisciplinary research, be accepting fusion of disciplines as new ones. Then, proper communication between disciplines is necessary. Finally, developing interdisciplinary research is extremely challenging as it often requires extra effort from an apprentice.
Are all the disciplines independent units nowadays?
Can we do molecular biology without technical, mathematical and computational support? Can we study cognitive science without knowledge of biology, physics, and psychology? Can we advance medicine without basic research in biology, physiology, electronics?
Bioinformatics and/or computational biology is an compelling case. Working in this field is being between biology, medicine, computer science, mathematics and statistics, the role of a computational biologist is sometimes reduced to a service. A biological lab may need a computational biologist to perform an analysis, restructure the data, that is needed for the biological discovery. Often, there is not enough space for research in computational biology itself, where the discovery does not depend on the original data but tools and approaches to complex, data-intensive biological problems. It may also happen the other way round when a computational biologist asks a bench researcher to perform an experiment to prove his theoretical model. In both cases, the long-term interdisciplinary partnership would probably fail. Wet and dry researchers should collaborate as equal with important research advances on both sides to assure a long-term equilibrium.
How did interdisciplinarity change over the years? Are all disciplines affected equally?
From the chart (Fig. .), we can notice that Social Studies of Medicine seems to be the most interdisciplinary field. In general Biology, Health and Biomedical Sciences seem to be more open into a flow of knowledge from other fields than humanities. On the extreme opposite of health, Clinical Medicine appears to be a very conservative field.
Figure .: Interdisciplinarity of different fields. “From 1950-2014, a field’s position is determined by how much its papers cite outside disciplines (x-axis), and by how much outside disciplines subsequently cite its papers (y-axis). (Some years, certain fields have too few references to be plotted.)”. Reprinted by permission from Springer Nature (Van Noorden 2015) © 2015 Nature America, Inc. All rights reserved.
Strengths, Weaknesses Opportunities, Threats (SWOT) of an interdisciplinary Ph.D. - personal perspective
I’m not good enough to do well something I dislike. In fact, I find it hard enough to do well something that I like — Jim Watson, Succeeding In Science: Some Rules Of Thumb (Csermely, Korlevic, and Sulyok 2007)
Being formed first in a double major in biology and mathematics, then participating in interdisciplinary research projects during my master studies, I can witness that the learning curve of multiple disciplines can be steep. It is also often associated with the frustration of not going deep enough in all of the disciplines or the feeling of being overwhelmed by the amount of knowledge.
Coming with the expertise in biology and mathematics, I got fascinated by complex biological systems. One way of study high-dimensional data is to reduce them into smaller interpretable units. This is what I tempted to achieve in this thesis in order to enrich our knowledge about tumor microenvironment and possibly contribute to orienting future research on immunotherapies.
However, being an interdisciplinary researcher was not always a privilege. To which category do I belong? To whom should I present my work? I often asked myself these questions. I also often encountered lack of understanding where my methodological results were not bringing enough of biological insights. Or the constraints of my biological application seemed very obscured and complicated for mathematicians, and my work often lacked important methodological advances.
Does it mean that my work is not accurate, useless? Probably, for many, it is not enough. However, I still hope that our findings will be interesting to some. I enjoy working with data and statistics that serve an actual purpose. The Tab. . summarizes Strengths, Weaknesses, Opportunities, and Threats (SWOT analysis) of an interdisciplinary project, in the way I perceive it.
| Strengths (internal, positive) | Weaknesses (internal, negative) | Opportunities (external, positive) | Threats (external, negative) |
| Having a holistic view of the problem | Not seeing details of the problem | Mulitple possibilities to convey research | Spending too much time filling knowledge gap |
| Being supervised by multiple experts | Following multiple, sometimes contradictory, advice on the same problem | Take advantage of synergistic effect of fields | Inhibiting effect of oppinions from different fields |
| Joining expertises of different fields | Not covering in details all the disciplines | Doing a new discovery | Obtaining too generic results |
| Using new/non standard approach | Experiencing steep learning curve | Raising interest in different expert domains | Not mastering the specific vocabulary of different fields |
| Having better understanding of complex processes | Being in constant need of help of domain experts | Making progress | Not being understood |
| Higher creativity | Creating a new field | Being hard to classify/ fall into a category | |
| Having great flexibility | Sovling many problems impossible to solve with traditional approach | Being considered as superficial | |
| Feeling a thrill of adventure | |||
| Being open |
Besides conducting research that crosses the boundaries of one discipline, I also could meet and work with inspiring people coping like me with filling the gap in understanding of interdisciplinary work, multiple supervisors and report to many institutions. I gained (even if only superficial) understanding of many topics in mathematics, statistics, data science, immunology, cancer but also oral and written presentation skills, time and work management
Is my thesis genuinely interdisciplinary? Does biology profits from mathematics and mathematics from biology? I will let you judge it.
What impact had biology on the statistical/mathematical modeling? The practical problems, systems that go beyond theoretical formulations challenge the theoretical tools. In my work, I did my best to fuse theory and practice that should serve a biological application. I can image the project more complete if the results of my work would inspire changes in biological experiments, uncover new paths to follow for experimental biologists or translational researchers.
The origins of the Ph.D. topic
The universe will lead me where I need to go. I am like a leaf in the stream of creation — Dirk Gently, Holistic detective
When finishing my master, I was looking for an interdisciplinary topic where I could deepen my quantitative skills and apply to a real-life healthcare problem. I came across a project proposed by Andrei Zinovyev in close collaboration with Vassili Soumelis. I was quite anxious that my knowledge of cancer immunology would not be sufficient to lead the project to a success. I recognize that the immune systems are very complex and dynamic system and many years of expertise are needed to grasp an understanding of it really. I had a great chance to work hand in hand with domain experts that would suggest me the direction I should take in my research.
The project started by causal exploration of different blind source separation or dimension reduction techniques and their ability to dissect bulk transcriptomic data into cell type-related units. We also faced a vital problem of lack of gold standard validation data that would define efficiency and accuracy of different methods.
I have spent fruitless efforts working on a bulk transcriptomic data simulation framework, important statistical issues come our way and probably another few years of a different Ph.D. would be necessary to solve them. In the meantime, many tools dissecting tumor bulk transcriptome were published. Serving a similar purpose, they used different means and assumptions, which left a space for my project to continue. In my third year, I am finally publishing a tool that performs the analysis I developed together with the Sysbio team members, and I can apply it to a corpus of publicly available data to learn about the actual question: the immune system infiltrating cancers and the context-dependent signatures (see Chapters 4 & 5).
In a parallel project, I worked on an exploration of a brand new data type: single cell transcriptomic (RNAseq) in the context of tumor microenvironment (see Chapter 6).
We have also participated in the Dream Idea Challenge, a project that aimed to put closer experimental and theoretical researchers (Annexe 1, (Azencott et al. 2017)) .
I have collaborated in numerous projects within and outside my team. Some of the projects resulted in publications, such as my work on analyzing pDC subsets of breast cancer Annexe2. Some others are in still preparation.
I have attended nine national, and international conferences, where I presented posters, gave talks and I got awarded with distinctions for my work.
Alongside with pursuing the compelling scientific research, I completed a wide variety of courses and I was teaching IT, Statistics and Mathematics at pharmacology faculty. Thanks to this extensive (>300 hours) training over three years, I am equipped with soft skills that not only helped me to shape my thesis project on the go but also, I hope, will help me to succeed in my future career path.
References
“The CRI | Centre for Research and Interdisciplinarity.” 2018. Accessed April 30. https://cri-paris.org/the-cri/.
Facilitating Interdisciplinary Research. 2004. Washington, D.C.: National Academies Press. doi:10.17226/11153.
Slavicek, Gregor. 2012. “Interdisciplinary -A Historical Reflection.” Int. J. Humanit. Soc. Sci. 2 (20). http://www.ijhssnet.com/journals/Vol{\_}2{\_}No{\_}20{\_}Special{\_}Issue{\_}October{\_}2012/10.pdf.
Ledford, Heidi. 2015. “How to solve the world’s biggest problems.” Nature 525 (7569): 308–11. doi:10.1038/525308a.
Van Noorden, Richard. 2015. “Interdisciplinary research by the numbers.” Nature 525 (7569): 306–7. doi:10.1038/525306a.
Csermely, Peter., Korado. Korlevic, and Katalin. Sulyok. 2007. Science education : models and networking of student research training under 21. IOS Press.
Azencott, Chloé-Agathe, Tero Aittokallio, Sushmita Roy, Ankit Agrawal, Tero Aittokallio, Chloé-Agathe Azencott, Emmanuel Barillot, et al. 2017. “The inconvenience of data of convenience: computational research beyond post-mortem analyses.” Nat. Methods 14 (10). Nature Publishing Group: 937–38. doi:10.1038/nmeth.4457.