Next “Meet the Experts” Workshop with Prof. Ronald Rutowski on April 13th, 2015 : Be welcome!

The BEE Doctoral school and the Biodiversity Research Centre at UCL are organising the next “Meet the experts” workshop with Prof. Ronald Rutowski (from Arizona State University, USA) as invited expert on :

“Sexual selection, Darwin, and Butterflies: Then and Now”

Scan 3A

Butterflies figured prominently in Darwin’s development of sexual selection theory.  This theory has proven to be a powerful and productive driver of research and valuable for understanding the evolution of many aspects of animal reproductive behavior.  However, its relevance to our understanding of butterfly coloration and other aspects of their reproductive biology was a point of contention in Darwin’s time, and even today empirical evidence for how sexual selection in butterflies plays out in nature, as opposed to the lab, is scant.   This talk will review some of the challenges and still unresolved issues that that we face in evaluating the nature and importance of sexual selection in butterflies.

When? April 17th, 2015, 14:00‐16:00

Where? Université catholique de Louvain, Place Croix du Sud, Carnoy Building, Room B059

Phd-students, post doctoral researchers and researchers  are all welcome. Participation is free, but registration is mandatory (by email

Incentive for phD-students: Participation to the workshop will count for 1 credit.

The organisers,

Camille Turlure & Caroline Nieberding


 More info:

Posted in Uncategorized | Leave a comment

Useful packages when working with SAS and R

This is a short post to share my experience for making the transition from SAS to R. There are dedicated resources about this transition such as books and several blogs here, here and here. I just would like to highlight two R packages that, I believe, are particularly useful when you come from SAS and start scripting in R, namely the sas7bdat package to directly access SAS datasets from R and the sqldf package for data management and number crunching.

When I started my PhD I learned SAS because there is a long tradition of using SAS in my research group with many preexisting scripts to automate steps such cleaning and processing of capture histories from mark-release recapture studies performed by my colleagues, or the automated reading and processing of the multitude of results files we get from our image analysis workflow. However, I was always a bit jealous of the nice plotting options that you have in R, the programming capabilities (I always found SAS macro somewhat clumsy) and the multitude of statistical analysis techniques which are already available or get implemented quickly. So I started to learn R about 1.5 years ago and use it more and more. I still use SAS quite a bit due to the above mentioned legacy of scripts and the ease and efficiency when dealing with big datasets (where R gets in trouble because they do not fit into memory), while analysis and plotting are done mostly in R now.

To efficiently combine SAS and R I found the sasb7dat package very handy. It allows to open the original SAS datasets without exporting them to a SAS xport library or converting them into a csv file which you could read in R. There are packages for input/output between R and other statistical software but these also rely on intermediate files (at least in the case of SAS). The package is still experimental but I have not observed any strange behaviour so far. The syntax is very simple, just as the read.csv() function:

data = read.sas7bdat(file="c:/data.sas7bdat")

Besides saving you some steps when exporting your data, you also have the advantage of working on just one dataset and not having several similar datasets floating around where you might even have forgotten, which one is the most recent etc.

My second recommendation is the sqldf package which is very powerful for all sorts of joining, sub-setting and aggregation of data. In R you would usually use indexing to access specific observations or columns but sqldf allows to stick to a familiar syntax for these tasks. There is quite a bit of good documentation available here. To illustrate the similarity of syntax here a simple demonstration how to create a subset a of another dataset in R and SAS:

treatment1 <- sqldf("select * from df where treatment = 1")

And now the same task in SAS:

proc sql;
create table treatment1 as 
select * 
from df
where treatment = 1;

These examples are fairly basic, but it is just intended to show how easily you may switch from SAS to R without the need of learning much new syntax before. If you are interested in the awesome ggplot2 package for visualization, you just need to install the sasb7dat and sqldf from CRAN to get your data into R and maybe rearranging it with sqldf and then you can start experimenting with ggplot2. Another great package worth checking out is knitr which allows you to create reports with embedded R code, plots and conclusions which is great for data exploration and keeping track of one’s conclusions.

Posted in data management, R, SAS | Tagged , | Leave a comment

Next “Meet the expert” workshop at UCL: Prof. Michel Loreau – November 5 th , 2013,

The BEE Doctoral school and the Biodiversity Research Centre at UCL are organising the next “Meet the experts” workshop with Prof. Michel Loreau (from CNRS, France) as invited expert on :

“Linking biodiversity and ecosystems: towards a unifying ecological theory”

M Loreau

When? November 5th, 2013, 14:00‐16:00

Where? Université catholique de Louvain, Place Croix du Sud

Phd-students, post doctoral researchers and researchers  are all welcome. Participation is free, but registration is mandatory (by email to and

 In addition, we seek highly motivated phD-students and postdocs to prepare the discussions before the workshop. If you are interested, please fill the doodle poll in order to find a suitable date for all of us. ( Do not forget to mention your email adress.

Incentive for phD-students: Active participation to the workshop preparation will count for 1 credit (ECTS) – Participation to the workshop only will count for 1 credit (ECTS)

The organisers,

Camille Turlure & Caroline Nieberding


Flyer: Meet the experts – M Loreau

Posted in Uncategorized | Leave a comment

Back from the course on Bayesian population analysis

Two weeks ago I attended the course “Bayesian population analysis using WinBUGS and JAGS”, given by Marc Kery and Michael Schaub in Sempach, at the Swiss Ornithological Institute. The course was based on the book by Marc Kery and Michael Schaub “Bayesian population analysis using WinBUGS: A hierarchical perspective”. I was impressed how Marc and Michael managed to cover basically the whole book within 5 days, starting with the implementation of mixed models in WinBUGS, and finishing with some examples of Integrated Population Models. Very informative lectures were complemented by solving at least one exercise per topic, although I wish there was more time for such exercises, because I usually learn more from trying to solve the practical questions. But, of course, it is not possible to have it all as the day has only 24 hours ;-)!

That was a wonderful course and a great opportunity to meet more people working in the same research field! If you are interested in taking the course, the next workshops will take place in Zuerich ( and Kuala Lumpur (, visit PHIDOT web site for more information.

Posted in Programming | Tagged , , , | 2 Comments

Causes of the spectral colour of population dynamics

The spectral colour of population dynamics and its underlying causes have received a fair amount of attention in the literature (e.g. García-Carreras & Reuman 2011; Inchausti & Halley 2002, 2003; Laakso et al. 2003). Most of the studies have demonstrated that the population dynamics of most species is reddened (Inchausti & Halley 2002; Laakso et al. 2003 but see Erb et al. 2001). Yet, no clear agreement exists about the drivers of this phenomenon. According to “reddened forcing hypothesis”, the population dynamics is following the colour of its environment, but there is a support both for (García-Carreras & Reuman 2011; Laakso et al. 2003) and against this explanation (Inchausti & Halley 2002). Alternative explanations have been proposed, just to name some of them: overlapping generations, consideration of spatial processes, etc. Erb et al. (2001) distinguished extrinsic (e.g. diseases, trophic-level interactions) and intrinsic (behavioural, such as dispersal or sociality) forces that are responsible for the “colour pattern” observed in population dynamics. However, stating that intrinsic factors are leading to direct density-dependent effects whereas extrinsic factors underlie the lagged effects is a bit of a simplification, as underlined by the authors themselves. Indeed, intrinsic processes, such as dispersal can also lead to lagged population responses.

It seems that after many years of studying the colouring of population dynamics, the analysis of long-term population time-series has so far uncovered the pattern, but the mechanisms leading to the observed colour of population dynamics remains unrevealed.

I can see two alternative ways to try to understand the mechanisms and distinct processes leading to specific colours of population dynamics:

– Mechanistic modelling of each process explicitly which would allow comparison of the issuing population dynamics with the pattern observed in the long-term time-series data. However, this approach is not free of drawbacks: the results of any model will largely depend on its structure and chosen parameters…

– Experimental testing of the alternative hypotheses about diverse forces driving the population dynamics and leading to its colour. Such designed experiments are not likely to be possible with large animals as study species, and here the long-established laboratory species models (such as daphnia, drosophila and tetrahymena, see about studies on this species in our lab on our blog) would be of great help! (see more on the choice of the study system here).

It’s hard to imagine that there is a clear-cut between intrinsic and extrinsic factors, and it is rather unlikely that some of them are predominantly driving the population dynamics of certain species. More likely, they are acting together, and, possibly, even an interaction among certain population processes gives a rise to a specific population dynamics with its spectral colour. It would therefore be an exciting research adventure: to use established laboratory systems to try with the well-designed experiments to disentangle the role of intrinsic and extrinsic factors, and in a second step, contrast the results of those experiments with the analyses of long-time population series collected for large mammals and birds.

I am looking forward to your critics and opinions about how it would be possible to reveal what processes are leading to the observed spectral colour of population dynamics!


Erb, J., M. S. Boyce, and N. C. Stenseth. 2001. Population dynamics of large and small mammals. Oikos 92:3-12.

García-Carreras, B., and D. C. Reuman. 2011. An empirical link between the spectral colour of climate and the spectral colour of field populations in the context of climate change. Journal of Animal Ecology 80:1042-1048.

Inchausti, P., and J. Halley. 2002. The long-term temporal variability and spectral colour of animal populations. Evolutionary Ecology Research 4:1033-1048.

Inchausti, P., and J. Halley. 2003. On the relation between temporal variability and persistence time in animal populations. Journal of Animal Ecology 72:899-908.

Laakso, J., K. Löytynoja, and V. Kaitala. 2003. Environmental noise and population dynamics of the ciliated protozoa Tetrahymena thermophila in aquatic microcosms. Oikos 102:663-671.

Posted in Uncategorized | Tagged , , | Leave a comment

Automated image analysis in ecology and evolution

Nowadays many ecologists and evolutionary biologists rely on experimental laboratory systems as a complement to field studies to test hypotheses under carefully controlled and standardized conditions.  Due to obvious size constraints of the laboratory environment model organisms are often microbes, insects or other arthropods (such as Daphnia sp.) but also small vertebrates such as fish or amphibians, which are small enough for easy manipulation and have short generation times to study longer term responses on the scale of weeks or months. These systems allow to obtain relatively long time series of population dynamics and detect evolutionary responses more easily than in the field and yielded many valuable insights.

Although laboratory systems have many advantages compared to cumbersome field work, they nevertheless require substantial amounts of time and manpower spent counting individuals and measuring phenotypes on the microscope. These tasks are often repetitive and tedious, so people get quickly tired and errors are likely to increase with each hour on the microscope. So why not let the computer do the work? Indeed, instead of observing by eye, images can be taken at low cost nowadays and then fed to the computer to do the counts and measurements. The field of computer vision and image processing has developed very quickly in recent years, just think about the omnipresence of QR codes all around to for example verify your train ticket or quickly direct you to a website when you scan the code with your smart phone. The potential of digital image processing and analysis in the biological sciences was already recognized about an decade ago and several researchers demonstrated later on its suitability in laboratory systems, but somehow the technique never really took a grip in the community.

In the paper “Implementing image analysis in laboratory-based experimental systems for ecology and evolution: a hands-on guide” which just got published in Methods in Ecology and Evolution, we argue that this is likely due to the daunting task of setting up such a system. Although the principles of image processing and analysis are sometimes surprisingly simple, setting up a workflow which is automated and validated to deal with a huge number of images is not always straightforward. So the paper tries to fill the gap of previous appeals by providing the necessary background on image analysis principles (e.g. different techniques for image segmentation), a guideline how to set up and validate each step of the workflow and some examples to illustrate its uses and advantages compared to manual approaches. Importantly, we do not leave the reader alone with figuring out the technical implementation, but provide ready-to-use scripts for image analysis for three commonly used and freely available open source software: ImageJ, Python and R. Each solution has specific strengths and weaknesses so it is up to the reader to choose the most appropriate solution for his system. We also provide a set of test images on which the scripts can be tested (see an example below).

Test image of a Tetrahymena thermophila culture where identified cells are outlined in yellow and enumerated.

Test image of a Tetrahymena thermophila culture where identified cells are outlined in yellow and enumerated.

Check out the paper and scripts to see whether the methodology cannot help you with automating parts of your research!

Posted in image analysis, microcosm, Programming | Tagged , , , | 2 Comments

Next “Earth and Life Talks”: Cancer as an ecological and evolutionary process by Pr. Michael Hochberg,

Université catholique de Louvain

14th of February 2013 –  Auditoire SUD 11, from 12h50 to 13h55

“Cancer as an ecological and evolutionary process”

Michael Hochberg

Research Director, Institute of Evolutionary Sciences, University of Montpellier II & External Professor at the Santa Fe Institute

Since the mid’ 1970s, cancer has been described as a process of Darwinian evolution, with somatic cellular selection and evolution being the fundamental processes leading to malignancy and its many manifestations (neo angiogenesis, evasion of the immune system, metastasis, and resistance to therapies). Historically, little attention has been placed on applications of ecology and evolutionary biology to understanding and controlling neoplastic progression and to prevent therapeutic failures.

This is now beginning to change, and there is a growing international interest in the interface between cancer and ecological and evolutionary theories. The objective of this talk is first to describe the basic ideas and concepts of cancer. I then present major fronts where the ecological and evolutionary perspective is most developed, in particular, that cancer generates substantial levels of genetic diversity, of which only a very small part is associated with tumor progression and metastasis. I discuss several of the most promising challenges and future prospects in fundamental cancer research.
Posted in Uncategorized | Leave a comment