This page is under development; more information will be added over time. For
now, it presents an overview of my programming work, and the projects I did
during my biology master at the Vrije Universiteit Amsterdam.
If you have any comments regarding this site, you can send an e-mail to
"INCA: INsect Count Analyzer"
Free-lance project for the Urban Wildlands
Group, Los Angeles, December 2001 – March 2002
Description: INCA is a user-friendly computer program that analyzes data
from transect counts: the number of insects along a fixed route on a number of
days. In its analysis, INCA uses the Zonneveld model, which contains four
parameters: an index of population size, the time of peak emergence, spread in
emergence times, and the death rate. INCA returns estimates for each of these
parameters, complete with standard deviations. In addition, it displays a chart
containing both your data and the model fit; this allows you to judge the model
INCA is freely available and can be downloaded from
the Urban Wildlands Group site.
Free-lance project, suggested by the Department of Theoretical biology, Vrije
Universiteit Amsterdam, August 2004 - present
Description: PlotReader is a plot digitizer. It is a Windows program
that enables you to retrieve data point coordinates from an image (scanned,
photographed or other) that contains a plot.
You simply open or paste a plot image in PlotReader, set 2 calibration points
per axis, and click all points you want coordinates for. PlotReader will then
calculate the original plot coordinates, and allows you to export these to file
PlotReader is freely available and can be downloaded from my site.
"Modeling Emiliania huxleyi: photosynthesis, calcification, and the
global CO2 increase"
Internship, department of Theoretical biology, Vrije Universiteit Amsterdam,
February 2000 – July 2001
Supervision: Dr. Cor Zonneveld, Prof. Dr. S.A.L.M. Kooijman
Summary: Coccolithophorids are a group of unicellular marine algae
considered to be responsible for the major part of the earth’s calcite
production. This process plays an important role in the global carbon cycle,
and may to a certain extent affect future developments in atmospheric CO2.
Our aim was to model the physiology of one of the most common
coccolithophorids: Emiliania huxleyi. The model should allow for application in
research related to the global climate and carbon cycle. Using the Dynamic
Energy Budget modeling approach (Kooijman, 2000), a dynamic model was
constructed that described all major carbon fluxes in a population of E.
huxleyi, as a function of external light intensity, CO2 , HCO3-
and NO3-. Steady state analysis of this initial model
revealed some serious model shortcomings related to the interaction between
calcification and photosynthesis. The model was modified to include a
biochemically more realistic representation of these processes: an internal CO2
pool was added, which supplied CO2 to photosynthesis, and obtained
CO2 from calcification. This corresponds more closely to the current
hypotheses regarding calcification. Steady state analysis showed obvious
improvements in the behavior of the model. Subsequently, the model was used to
describe the behavior of Emiliania at various concentrations of ambient CO2,
illustrating the model’s use in its targeted area of research. Although the
model was able to describe the data, we must conclude that the datasets were
too limited to constrain the parameter values. Clearly, the complexity of the
model places severe requirements on the quality and quantity of the data used.
If these requirements were satisfied, the model could be valuable not only in
research related to the global climate, but also to that focusing on the
biochemistry and physiology of calcification.
"Mining MIM: Phenotype clustering as source of candidate genes"
Internship, department of Bioinformatics, Wageningen University, November 2002
- April 2003
Supervision: Prof. Dr. Jack Leunissen
Summary: Given a description of an anomalous phenotype, one may well be
able find the gene(s) responsible for the anomaly in a list of genes associated
with very similar phenotypes. Thus, a technique that identifies similar
phenotypes could be exploited as a resource of candidate genes. In this
project, we aim to develop such a technique.
As test case, we take the Online Mendelian Inheritance in Man (OMIM) database,
which contains 14,000+ articles describing human inheritable traits and
diseases. To determine the similarity between OMIM articles, we first
characterize each article by the occurrence of a select set of (bio)medical
terms. This set comprises the ‘anatomy’ and ‘disease’ categories of the Medical
Subject Headings (MeSH) thesaurus. All MeSH entries are regarded as potential
features of an OMIM article; per article, the value of any such feature is
taken equal to the number of occurrences of the corresponding entry. In
addition, we add for each matched entry its MeSH ancestors (i.e. the entries
describing a superset of the entry) to the feature vector. This allows for
closely related terms (sharing ancestors) to contribute to between-article
similarity rather than diminishing it; in effect, the system has become
term-relation-aware. Finally, feature vectors are corpus-size normalized, and
feature values are weighed according to a scheme common in the field of
To determine similarity between phenotype feature vectors, we calculate their
length-normalized correlation (i.e. the cosine of the angle between the
vectors). With this similarity measure, we can list phenotypes similar to a
given reference, ordered by proximity (i.e. ranking). In addition, we perform
hierarchical clustering (UPGMA) on the full set of phenotypes.
To establish whether phenotype similarities indicate genotype similarities, we
derive a genotype feature matrix for a set of ±1,000 OMIM articles, selecting
as features the Gene Ontology (GO) terms indirectly associated with the
articles. Subsequently, we calculate the correlation between phenotype- and
genotype proximity matrices, both pre- and post-clustering.
Result evaluation is difficult, as the set of phenotypes is large, and
calculation of the significance of phenotype-genotype correlation coefficients
is not feasible. However, both random sampling of article neighbors and
qualitative evaluation of correlation coefficients indicate our system could
prove a valuable resource of candidate genes. This is specifically the case for
human phenotypes (i.e. OMIM); current application for other species may be
troublesome due to low-quality phenotype descriptions, and lack of a dictionary
with phenotype-describing terms.
"Religion: Spandrel of the Brain? The Evolutionary Psychology of Religion"
Literature study, department of Theoretical biology, Amsterdam, July -
Supervision: Dr. Cor Zonneveld
Summary: During the last decades, research into the biological origin of
religion has gained considerable momentum. Religiousness has been shown to be
heritable – though only to a moderate extent – which has prompted various
scholars (e.g. D.S. Wilson, R.D. Alexander) to argue religion is an adaptation:
a genetically inherited trait that increases fitness. However, as Stephen J
Gould has forcefully argued, not all inheritable traits need be adaptations; in
various ways traits can emerge and persist in evolution without increasing
fitness per se. In this literature study, I explore the possibility that
religion is not an adaptation, but exists for different reasons.
Non-adaptive traits can persist as by-products of adaptive traits. A number of
scholars (P. Boyer, L. Kirkpatrick, S. Atran, S.J. Mithen, S. Guthrie) have
argued religion qualifies as a by-product of cognitive hardware. In their view,
religious thought and behavior is produced by various brain mechanisms that in
themselves are adaptive. These were preserved in evolution because they solved
particular (non-religious) problems faced by our ancestors. By-product theories
of religion provide an interesting and plausible alternative to their
adaptationist counterparts: they are capable of explaining the full spectrum of
religious manifestations, without invoking any function for which they were
Five million years ago, the evolutionary paths of man and chimpanzee separated:
our last common ancestor eventually gave rise to a distinct human ancestor – Australopithecus
– and an ‘ancestral’ chimpanzee. In the period to follow, the ape-like Australopithecus
came to evolve into the modern Homo sapiens sapiens; selection pressures
over the last five millions years gave rise to all traits commonly considered
uniquely human (e.g. language, culture). What were these selection pressures
like? Many have argued that our ancestor most prominently was under great
pressure to organize: formation of ever-larger groups would have been
indispensable for coordinated hunting- and gathering expeditions, and defense
against predators (animals, but also other human groups).
To enable increases in group size, our brain has become strongly focused on
other humans: it ascribes them beliefs, intentions and emotional states, feels
their pain, shares their moral codes, and watches closely if others adhere to
such codes. In fact, the human brain has become so strongly focused on humans
that it quite easily perceives humans when there are none: any inexplicable
phenomenon, any single or remarkable event is ascribed to human agents. In
addition, the new cognitive hardware for active reasoning about other humans
tends to over-activity: it fails to shutdown when observing familiar corpses
(thus leaving the brain with the concept of a bodiless human agent), and
prefers to represent the subconscious self as beliefs and intentions of others.
As a result, the brain cannot help but observe an environment flooded with
(signs of) powerful, invisible supernatural agents. Given the means through
which it does so, it is not surprising that the brain holds quite elaborate
assumptions on the nature of such agents: they are thought powerful, invisible
or acting remotely, and involved in social exchange (and hence, to a certain
extent manageable – one could achieve something by engaging in social
interaction with the supernatural).
Summarizing, the brain inadvertently develops the notion of relevant
supernatural agents: certainly an ideal starting point for religious thought
and practices. Five million years 3 of specializing in social interaction
rendered a well-adapted brain that fluidly – yet inadvertently – constructs its
last update: November 12, 2010. Copyright Jorn Bruggeman.