Decentralised self-regulation of a hydroid colony in the course of growth
N.N. Marfin
Department of Invertebrate Zoology, Lomonosov Moscow State University
Colonial hydroids can be used as a highly informative model of self-organisation in biological systems when addressing at least two original tasks.
A hydroid colony is a single organism that is distinguished from unitary organisms by its modular structure, i.e. multiple replication of stereotypical parts (or organs). Unlike unitary organisms, modular organisms lack organs of central regulation and coordination. Their nervous system is difuse: they have no nerve ganglia, let alone a brain.
Nevertheless, this «brainless» organism displays all attributes of high integration: 1) a single body, 2) a single body cavity, 3) an efficient transportation system (similar to the vascular system of more evolved animals), which provides the physiological integration of all zooids, stolons and stems into a single entity, 4) proportional ratio of body parts (zooids, growing regions, stems, stolons), 5) capacity to regulate the growth and branching of body parts depending on food supply to the whole colony, and not only to those parts of the colony that are adjacent to growing regions.
Experiments have demonstrated that in case of food shortage, the bulk of it is «delivered» to the most remote peripheral parts of the colony to the growing tips of stolons and new stems, regardless of where the food was captured by the «multimouth organism». Mathematical modelling can become the main research method for studying the limits of efficiency in self-regulation of proportional composition, the rate of the process, and its correlation with morphological parameters of the colony.
The transportation of hydroplasm is carried out through a transport system of the pulsatile peristaltic type. The body of the colony resembles a branching non-circular network of elastic tubes of the same diameter (coenosarc), which is filled with liquid (hydroplasm) and has no direct contact with the outside environment. All parts of the coenosarc can actively contract, but not actively expand. Expansion is caused by the pressure of hydroplasm flowing from adjacent parts of the colony, which undergo contraction of the coenosarc. Such flows are localised. There are either zooids or so-called «growing tips» at the ends of the coenosarc. Both pulse with higher amplitudes. The whole colony presents a network of tubes with pear-shaped organs at the ends.
When food enters the colony, local movements of hydroplasm is activated and affect each other, which results in mutual adjustment of separate pulsators (zooids and growing tips), and thus creates continuous flows of hydroplasm of higher rate and range. Such a system of decentralised self-regulation is equifinal, i.e. does not depend on the configuration of the colony and its transformations in the course of growth or under the impact of external factors. The talk lays down the fundamental principles of decentralised self-regulation.
Mathematical modelling helps to reveal the connection between the efficiency of the transport system (physiological integration of the colony) and morphological features of the colony structure (diameter of the coenosarc, zooids, growing tips, segment length, etc.).
References
Marfenin N.N., Kossevitch I.A. (2004) Morphogenetic evolution of hydroid colony pattern / Coelenterate Biology 2003: Trends in research on Cnidaria and Ctenophora / Eds. Fautin D.G., Westfall J.A., Cartwight P., Daly M., Wyttenbach C.R. // Hydrobiologia, 2004. – P.319-327.
Marfenin N.N. (2002). Decentralized self-regulation of an organism integrity // Jurnal obshey biologii (J.General Biology), 63, 1: 26-39 (in Russian).
Marfenin N.N. (1997) Adaptation capabilities of marine modular organisms // Hydrobiologia, 1997. Vol.355. P.153-158.
Biophysical models for ion dynamics in cell polarization and apical growth
Filipa Alves
Instituto Gulbenkian de Ciência, Oeiras, Portugal
Neurons, hyphae, root hairs and pollen tubes all share comparable mechanisms to regulate cell polarization, apical growth and chemotactic spatial orientation.
We are using the germinating pollen as a model system, as different experimental approaches have already produced high quality data, ranging from molecular biology to electrophysiology, imaging and transcriptomics. In these cells, ion dynamics plays a central role in the establishment of polarity and in the regulation of growth and chemotaxis.
Experimentally, we observe and quantify complex temporal and spatial patterns of the ion fluxes across the plasma membrane and the free ion concentrations in the cytoplasm. Based on the available data, we are developing computational models integrating ion fluxes and transporters distribution. These models predict the observed intracellular ion gradients, shedding light on the minimal necessary conditions to establish these gradients and providing some rationales for species-specific differences.
Evolution of Biomineralization Protein Domains in Metazoans
Bishoy Hanna
School of Natural Sciences, University of California, Merced
Biomineralization is a highly regulated process, in which various layers of complexity contribute to the final outcome of a mineralized skeleton. Multiple animal phyla gained this important feature during the Cambrian explosion. Due to the diverse range of skeletons found in modern day animals it has been difficult to formulate an evolutionary framework that explains the emergence of mineral formation in various metazoan lineages.
To test different hypotheses regarding the mechanisms involved in the evolution of mineralized skeletons, we have employed a multi-tier bioinformatics approach where we have identified the domains involved in biomineralization proteins and their interaction networks revealing the complex nature of this important biological process. Moreover, we are able to predict new proteins involved in biomineralization. The network analysis provided us with the ground to establish a “Pyramid of Complexity” framework to understand the evolution of biomineralization on the protein level on various levels of complexity.
Finally, we have established a pipeline (BiomMine) to find and analyze biomineralization proteins in current databases and from next-gen sequence data
Design principles of mammalian epigenetic gene regulation: Systems and synthetic biology approaches
Lisette C. M. Anink1, Diewertje G. E. Piebes1, Maartje C. Brink1, Andreas Kremer2 Anne Schwabe3, Frank J. Bruggeman3, and Pernette J. Verschure1
1Swammerdam Institute for Life Sciences/ Netherlands Institute for Systems Biology, University of Amsterdam, Amsterdam, The Netherlands 2Erasmus MC, Department of Bioinformatics, Dr. Molewaterplein 50, 3015 GE Rotterdam, The Netherlands 3Center for Mathematics and Computer Science/ Netherlands Institute for Systems Biology, Amsterdam, The Netherlands
Epigenetic gene regulation is crucial for cellular identity in higher eukaryotes. Derangements in epigenetic gene control have severe effects on cell behaviour and contribute to a diseased state. To understand the molecular mechanisms of functional genome organization in mammalian cells, we use a systems-, synthetic biology approach combining quantitative experiments with mathematical modeling. We follow two complementary approaches: 1) Genome-wide gene expression/epigenetic profiling and computational modeling to unravel epigenetic network behavior and 2) Analysis of rationally designed mammalian synthetic cell systems to mechanistically uncover in vivo design principles of epigenetic gene regulation.
Approach 1 concerns a collaboration with Dr. A. Kremer (ErasmusMC, Rotterdam, the Netherlands), studying the switch in phenotype caused by the neurodegenerative disease Huntington’s disease. Regarding approach 2, we created synthetic cell systems enabling (i) to toggle the epigenetic state of a reporter gene cassette integrated at predefined genomic regions (region of high and low gene expression), (ii) to systematically measure the causal relationships between chromatin structure, gene activity and the epigenetic state and (iii) to identify the parameters allowing for robust regulation while maintaining low molecular noise propagation. Computationally, we created coarse-grained models explaining stochastic gene activity and ‘spreading’ of the epigenetic nucleosomal state predicting the outcome of our biological experiments. Using chromatin immunoprecipation (ChIP) we show that histone modifications related to an ‘active’ chromatin state and RNA polymerase II are enriched at the cassette integrated in a high gene expressing genomic context. Targeting of HP1 shows a sequential kinetic increase in the reporter protein and transcript levels and a reduction in the levels of ‘active’ histone marks and increase of ‘repressive’ marks. This research is proof-of-principle to create more complex synthetic systems including multistability or oscillation opening an unexplored field of research with great potential for medicine.
Modelling protein protein interaction based on protein abundance data in neurons
Joachim Kutzera, Huub Hoefsloot, Age Smilde, Ka Wan Li, Guus Smit
This project is about the reconstruction of protein-protein interaction networks based on abundance data from immunoprecipitation.
Much research is done on exploring PPI (protein protein interaction), concentrating on two or three specic proteins in cells. Communication pathways were discovered bit by bit through combining the work on different proteins. Until today, it is diffcult to discover bigger protein communication networks in a cell, as it includes maybe thousands of proteins with different functions, and we cannot observe them interacting. The interaction ability itself can be detected for two proteins with tools like the Yeast-2-Hybrid system. However, this method cannot proof, whether two isolated proteins interact, it is only suitable to test on two gene sequences, whether the proteins resulting from the expression of the genes can connect to each other or not.
A widely used biochemical method for purifying specifc proteins is IP (immunoprecipitation). This method uses antibodies that stick to a specific protein, allowing to isolate this protein together with all other proteins that are currently connected to it. After identifcation and quantifcation of the purifed proteins, the result is a table with abundance values for this proteins in one single IP experiment. With the help of the abundance values, it is possible to get an insight into the protein-protein interacton network, especially, when a high number of IP experiments are done with different antibodies on the same cell sample to get different parts of the network.
The project is divided into two different parts. Part one concentrates on measuring protein abundance data (so called "wet work"), in part two we try to model the interaction networks from this data ("dry work"). My work is about the second task, as well as the presentation.
In the real experiment, cell material extracted from mouse neurons is analyzed with IP. The purified proteins are split into peptides through trypsin digestion and the peptides are measured using LC/MS with tandem-MS. The identification of the peptides and the quantification of the resulting proteins is done with help of peptide spectra databases.
The determination of protein abundance values can be simulated in silico for a theoretical network. This network is represented by a graph {V,E}, composed of nodes V for proteins and edges E between these nodes, that stand for the interactios.
Supposing, that one pulls on a node v in the network and each edge has a certain propability to break. Then, the propability for each node v* to be still connected directly or indirectly to v is below 1 and can be calculated. In the simulation, these connected nodes represent the proteins in one IP-experiment, that do not stick directly to the antibody, but get pulled out indirectly. This simulation is repeated for all nodes. A matrix containing the connectivity propability for all nodes to all other nodes is the result.
This matrix is correlated partly to the data from real IP experiments. However, the simulated networks are much smaller than the networks in the real measurements. One can easily pull on an arbitrary node in the simulation but not in the real IP, as there is no antibody available for every protein. For this and other reasons, the simulation strongly simplifies the biological processes, but seemes to be sufficient to explain the protein abundancies in IP results from a real protein network.
The presentations shows, that a constructed example communication network can be reconstructed roughly from the calculated connectivity matrix. The reconstruction starts with the detection of direct connections between nodes and continues with finding closely connected clusters. The connections between these clusters can be detected as well, but work on this is still in progress.
Towards an Understanding of Peptide-Mineral Interactions: a Predictive Approach
David .J. Belton, Marion Limo, Valeria Puddu, Rajesh Ramaswamy and Carole C. Perry
Interdisciplinary Biomedical Research Centre, School of Science and Technology, Nottingham Trent University, Clifton Lane, Nottingham, NG11 8NS
A goal of our research is to uncover and understand the interactions between (bio)molecules such as peptides and inorganic materials and to identify ‘rules’ or ‘guiding principles’ that could explain and, furthermore, predict structure and properties for a wide range of (bio)molecule-mineral systems.
At present we are studying a range of amorphous and crystalline materials to (a) investigate the ability of peptides to generate materials and in particular to explore and quantify the energetics of peptide binding to well characterised particles, (b) to complement the experimental studies with computational studies of the peptides, the materials and the interfacial interactions between the peptides and the materials, and (c) to perform studies aimed at understanding sequence diversity for a particular material, including the development of representational tools that can be readily used by technologists in the field.
In the current contribution we will start by looking at the important role that interfaces play in both biominerals and in the development of materials based technologies and consider the effect that amino acid for amino acid substitutions play in the function of a material using examples from biology and silicified sponge spicules as examples. We will then discuss how the study of mineral-peptide/ protein interactions can be adapted for materials that are not commonly found in nature. Using zinc oxide as an example we will describe how peptide composition affects the morphology of the mineral found and how both laboratory experiments and in silico experiments can help us to understand the interface between biomolecules and minerals. The presentation will conclude with questions that remain to be answered including the question as to the role of informatics in the prediction of interfacial interactions.
Coherent and autonomous modes of cell behavior in the gastrulation of cnidarians
Dr. Yulia Kraus
Moscow State University, Biological faculty, Dpt. of Evolutionary Biology
It is possible to distinguish at least two types of morphogenetic processes in embryogenesis using as a key feature the main mode of cell behavior. Morphogenesis of the first type goes on via changes in cell shape occurring in a coordinated fashion, directed by cell interactions based on signalling molecules or on the external mechanical forces acting on the epithelial cells (Leptin, Roth, 1994). Epithelial morphogenesis in the embryonic tissues of high metazoans is normally based upon cooperative cell-to-cell interactions (Belintsev et al., 1987). Endoderm invagination in echinoderms or neural tube formation during neurulation of amphibians are the best examples of such a morphogenesis. Changes in the shape of each individual cell affect the mechanical state of all neighboring cells and thereby provoke their active re-shaping. In this case, epithelial sheet morphogenesis can be considered as a side effect of a coherent self-reinforcing shaping of individual cells. Very often (but not always), epithelial cells undergoing the morphogenesis form a spatial unfolding - continuous spatial series of cell shapes corresponding to the succession of changes in the shape of a single cell (Cherdantsev, 2006). Morphogenesis of the second type is based on the active autonomous reshaping of single cells. In Drosophila gastrulation, prospective mesoderm cells constrict their apices and become bottle cells in an apparently stochastic order (Sweeton et al., 1991). It seems that all mesodermal cells independently follow their developmental program and independently undergo the cell shape changes (Leptin, Roth, 1994).
Cnidaria are the simple metazoan animals. A striking feature of this phylum is a great variety of morphogenetic processes that can be observed in the different cnidarian species during gastrulation. Gastrulation of cnidarians provides good examples of the both types of morphogenetic processes described above.
Formation of the epithelial ectoderm from non-epithelial cells is the major event in the gastrulation of the colonial hydroid Dynamena pumila (Kraus, 2006). This process is based upon cooperative cell-to-cell interactions. Each cell that joins forming epithelial sheet fragment gradually changes its shape in concert with the changes in the shape of neighboring cells and in the shape of an entire fragment.
The sea anemone Nematostella vectensis gastrulates by invagination of a pre-endodermal plate consisting of cells undergoing epithelial – mesenchymal transition (EMT) and acquiring the ?gbottle?h shape (Kraus, Technau. 2006). It seems that EMT spreads other the region of presumptive endoderm. The spreading of EMT is an extremely variable process: the embryos differ from each other in the geometry of pre - endodermal plate and in the number of cells involved. Thus, it is possible to conclude that this process is based upon cooperative cell-to-cell interactions.
Gastrulation in the colonial hydroid Clytia hemispaerica has been described as an ingression of the presumptive endoderm cells from the specialized oral territory established by the action of maternally localized determinants (Metschnikoff, 1886; Momose, Houliston, 2007). Single bottle cells or very small groups of bottle cells are stochastically interspersed among ordinary epithelial cells. It means that EMT never spreads over the entire oral territory, and formation of bottle cell in Clytia is an autonomous process.
In order to find general rules governing the morphogenetic movements based upon coherent and autonomous cell behavior, it would be very interesting to analyze a set of the developmental parameters using a biomechanical models of cnidarian gastrulation.
References
Belintsev B.N., Beloussov L.V., Zaraisky A.G., 1987. Model of pattern formation in epithelial morphogenesis. J. Theor. Biol. 129: 369-394.
Cherdantsev V.G., 2006. The dynamic geometry of mass cell movements in animal morphogenesis. Int. J. Dev. Biol. 50: 169-182.
Kraus Y., Technau U., 2006. Gastrulation in Nematostella vectensis occurs by invagination and immigration: an ultrastructural study. Dev. Genes Evol. 216: 119–132.
Kraus Yu. A., 2006. Morphomechanical programming of morphogenesis in cnidarian embryos. Int. J. Dev. Biol. 50: 267-275. Leptin M., Roth S., 1994. Autonomy and non-autonomy in Drosophila mesoderm determination and Morphogenesis. Development 120: 853-859.
Metschnikoff E., 1886. Embryologische studien an Medusen. Ein Beitrag zur genealogie der Primitiv-organe. Alfred Holder, Vienna
Momose T., Houliston E., 2007. Two oppositely localised Frizzled RNAs as axis determinants in a cnidarian embryo. PLoS Biol 5(4): e70. doi:10.1371 / journal.pbio. 0050070
Sweeton, D., Parks, S., Costa, M., Wieschaus, E.91. Gastrulation in Drosophila: the formation of the ventral furrow and posterior midgut invaginations. Development 112: 775-789.
Gene regulatory network inference by integrative evolutionary computation
Alina Sirbu, Heather J. Ruskin, Martin Crane
The advances in high throughput techniques for measuring gene expression levels have triggered considerable efforts in reverse engineering gene regulatory networks (GRNs). Among these, quantitative models have received much attention, but, to date, can not be reliably built for specific processes, due to limited time series data, (typically used for inference). These series are noisy and contain insu±cient time points, so that the inverse problem is under-determined, which leads to many possible model fits for the data. To reduce the under-determination problem, integration of other types of widely available biological data is a promising approach, and has not been fully explored, to date.
Here we present one such integrative effort based on evolutionary optimisation. This employs expression data from knockout experiments, pair-wise correlation between gene expression patterns, annotated transcription factors, binding site affinities and promoter sequences to drive the algorithm towards more promising regions of the possible network structure space, by means of a customised mutation operator. A composite evaluation criterion is also employed, that aims to reduce the issues related to noise overfitting and under-determination, which arise from use of MSE alone to validate models. The more sophisticated evaluation basis includes consideration of network structure, (employing binding site a±nities), correlation to initial data and error on noise added data.
The approach has been validated on both synthetic and real gene expression data, and effects are discussed here. On synthetic data, both quantitative (MSE-based) and qualitative (direct interactions) improvements have been obtained by using knockout ex- periments and pair-wise correlation alone, within the mutation operator, and without including structure information in the evaluation criterion. On real data, the inclusion of binding site affinity information in the evaluation criterion proved to be important in obtaining good fit on the test data and also led to inclusion of more direct interactions. Using MSE only for evaluation, together with the customised mutation operator, did not lead to improved results. Due to the nature of these data, models including direct inter- actions were not able to improve on simulation of the training time series, indicating that the MSE criterion for evaluation is too crude. However, by adding structure informa- tion from affinities, models were obtained, which included an increased number of direct interactions and with good simulation abilities.
Keywords: data integration, gene regulatory network, mathematical modelling, evo- lutionary computation, evaluation, mutation
An integrated approach to infer the gene network in early development of the cnidarian Nematostella vectensis.
Daniel Botman, Jaap A. Kaandorp
We have developed new methods to analyse spatio-temporal gene expression patterns (in situ hybridizations) and morphological data (based on confocal light microscopy images) of Nematostella during early embryonic development.. The gene expression images are processed with two-dimensional geometry extraction methods and cell-layer decomposition methods to consistently compare and model the expression patterns.
Moreover, a three-dimensional tool has been developed for geometry extraction and decomposition in order to model expression patterns that are not radially symmetric about the main body axis. We assume that embryo shape remains symmetric with regard to the oral-aboral axis, so the embryo's original geometry is applied to all dissections through this axis.
As a proof of principle, this procedure has been applied to infer the parameter values of a proposed transcription network for Nematostella primary axis formation. This network consists of nine suitable genes that are differentially expressed, and some other proposed transcription factors. Currently, we are looking into the gene regulation network of the early bmp pathway in Nematostella. Genes from the bmp cluster have been shown necessary for secondary (directive) axis formation.
Interspecies translation of disease networks increases robustness and predictive accuracy
Seyed Yahya Anvar1,*, Allan Tucker2, Andrea Venema1, Gert-Jan B van Ommen1, Silvere M van der Maarel1, Vered Raz1 , Peter AC ‘t Hoen1
1Center for Human and Clinical Genetics, Leiden University Medical Center, P.O. Box 9600, 2300 RC Leiden, the Netherlands. 2Center for Intelligent Data Analysis, School of Information Systems, Computing and Mathematics, Brunel University, Uxbridge, Middlesex, UB8 3PH, UK.
* Correspondence: s.y.anvar@lumc.nl
The derivation of gene regulatory networks from high-throughput expression data via machine learning strategies is problematic. Their reliability is often compromised by limited and highly variable samples, heterogeneity in transcript isoforms, noise, and other artifacts. Hence, a probabilistic approach is needed that can represent complex stochastic relationships, integrate different types of data, weight the reliability of the data, and accommodate noise and missing values. We developed a novel algorithm, dubbed Dandelion, in which we construct and train intraspecies Bayesian networks which are translated and examined on independent test sets from other species in a reiterative procedure. The interspecies disease networks are subjected to multi-layers of analysis and evaluation criteria which lead to the identification of the most consistent relationships within the network structure. We demonstrate the performance of our algorithms on datasets from animal models of oculopharyngeal muscular dystrophy (OPMD) and patient materials. We show that the interspecies proteasome networks provide highly accurate predictions on gene expression levels and disease outcome whilst the stability of these networks increases after interspecies translation. Unlike existing modeling approaches, our algorithms do not require notoriously difficult one-to-one mappings of protein orthologues or alternative transcripts and can deal with missing data. We show that the OPMD-association of potential key regulators can be reproduced and validated on an unseen and independent model system. This study presents a state-of-the-art strategy in constructing interspecies disease networks that provide crucial information on functional regulatory relationships among genes, leading to better understanding of the molecular mechanisms underlying the disease.
Correlation between Mitochondria Genomes Structure and Taxonomy
Michael G.Sadovsky
Institute of computational modelling of SD RAS
* Correspondence: msad@icm.krasn.ru
Mitochondria genomes are well known both for their conservatism, and relative simplicity. We studied a relation between a structure and the taxonomy of genome bearers. Frequency dictionary W3 of triplets makes a structure. Each genome is a point in 63-dimensional space; besides, three more spaces have been used to develop an unsupervised classiffication:
-{ 32-dimensional subspace of the triplets from the "left" part of a complimentary palindrome;
-{ 32-dimensional subspace of the triplets from the "right" part of a complimentary palindrome, and,
-{ 32-dimensional space of frequency differences of triplets making a complimentary palindrome.
Also, similar classiffications have been implemented at the relevant spaces of information values of triplets. To reveal the relation, an unsupervised classiffication has been implemented, in a frequency space, due to K-means technique. An elastic map clusterization has been implemented, as well. The composition of the classes and clusters yielded due to the classiffication (or clusterization) was studied. To develop a classiffication, the original genome database has been hashed to eliminate volatile elements, and equalize clade abundance.
Also, similar classiffications have been implemented at the relevant spaces of information values of triplets. To reveal the relation, an unsupervised classiffication has been implemented, in a frequency space, due to K-means technique. An elastic map clusterization has been implemented, as well. The composition of the classes and clusters yielded due to the classiffication (or clusterization) was studied. To develop a classiffication, the original genome database has been hashed to eliminate volatile elements, and equalize clade abundance.
Fig. 1 shows the distribution of 1132 genomes in various spaces; subfigures (a) and (b) show elastic maps (both are rigid, and 12 x 12 in size) developed in two different spaces: the former is developed at
(a) 63-dim. space, frequency; 3 classes.
(b) 32-dime. space, values; 3 classes.
(c) Clades distribution.
63-dimensional space of triplet frequencies, and the latter is developed at 32-dimensional space of the differ- ences of information values of triplets composing complimentary palindromes. Here three classes obtained due to K-mean classiffication are shown, in colour. Finally, subfigure (c) being identical to subfigure (a) from the point of view of the space structure, shows the clade distribution, with no classiffication. Here five most abundant clades are shown: fishes are in orange, amphibians are in light green, mammalians are in light blue, insects are yellow, and in dark green are shown dinosaurs and lepidosaurs. All classiffication were developed with ViDaExpert software [1].
Strong correlation between clades, classes developed due to K-means classiffication, and clusterization yielded by elastic maps is observed. Moreover, further clusterization is observed, for subgroups of genomes belonging to a single clade. Both biological, and statistical (combinatorial) issues standing behind the patterns shown above are discussed.
References
[1] Zinovyev A., http://bioinfo-out.curie.fr/projects/vidaexpert/
Models in Information Transformation in Genomes
Anna A. Koval *, Michael G. Sadovsky **
*Siberian Federal University, anya.a.koval@gmail.com **Institute of computational modelling SB RAS, msad@icm.krasn.ru
A study of statistical properties of nucleotide sequences allows to reveal some new issues in fundamental problems of molecular biology, genomics and evolution theory. Frequency dictionaries are often used to determine the properties and features of DNA sequences. This approach yields a powerful tool to investigate both statistical, and information properties of nucleotide sequences. Here we present some models in information transformation in genomes manifested in information capacity behaviour.
Information capacity is a measure of deviation of the reconstructed dictionary from the real one, or informally speaking, a measure of "unpredictedness" of various words in the dictionary. It is determined through the comparison of real frequency of words with their expected frequency obtained from the consideration of shorter strings. Since the information capacity is defined through the mutual entropy of the real frequency dictionary against the reconstructed one, then:
(1) where ~Sq is the mutual entropy of the frequency dictionary (of the thickness q), and Sq is the absolute entropy of the frequency dictionary.
We studied whether the information capacity exhibits some unusual properties: it differs for real DNA sequences as compared with various model ones. To verify this assumption, we compared the information capacity of the real DNA sequence to that latter determined for model DNA sequences. Model DNA sequences have the same structure in non-coding regions, but the frequencies of synonymous codons in coding regions have been replaced as follows: Firstly, with mean values; secondly, the strongest bias due to the change of all synonymous codons for the rarest one; and thirdly, the change for the most frequent one, within a group. All three types of model sequences kept the same transcription, as at the original sequence. Also a number of other patterns of randomization will be implemented, as well.
Structured Therapeutic Interruptions of HAART optimized by simulated annealing
Emiliano Mancini, Filippo Castiglione, Massimo Bernaschi, Peter Sloot
Highly Active Anti Retroviral Therapies(HAART) are currently the only therapies for treatment of HIV infections, extending the life expectancy of HIV positive individuals. Unfortunately such therapies have several issues that force therapy interruption including the emergence of drug resistant mutants and side effects of the drugs. Early experiments on Structured Therapeutic Interruptions (STI) seemed to yield good results on the short term but recent studies showed a negative effect of STI on the long term. Although the use of STI is still very controversial, tests on humans have slowed down because of the high risks.
In this scope a computational model to investigate the effect of STI on the infection dynamics seems extremely useful. The HIV infection and HAART are simulated with C-ImmSim, an agent based model of the immune system that reproduces the immune response of different virtual patients to the HIV-1 infection. We use a Simulated Annealing algorithm to search for the optimal schedule that maximizes immune restoration and minimizes both the viral load and the dose of drugs administered to the virtual patients.
Individual based modeling of coral colony growth.
Maxim V. Filatov *, Jaap A. Kaandorp *,
* Section Computational Science, Faculty of Science, University of Amsterdam Science Park 904 1098 XH Amsterdam The Netherlands
Corresponding author email: M.V.Filatov@uva.nl
In previous work [1] we have developed a computational model of the growth of a scleractinian coral colony. Using our model we show how environmental factors such as nutrient distribution and light availability affect growth patterns of coral colonies. To compare the simulated coral growth forms with those of real coral colonies, we quantitatively compare our modeling results with coral colonies of the morphologically variable Caribbean coral genus Madracis [2]. Our results show that simulated coral morphologies share several morphological features with real coral colonies (M. mirabilis, M. decactis, M. formosa and M. carmabi). Our model is able to partly capture the morphological variation in closely related and morphologically variable coral species of the genus Madracis. To improve this model we use an individual based modeling approach to control the growth of the colony.
In this approach the behavior of individual polyps is modeled. This new model includes the following features: (i) the uptake of dissolved inorganic carbon can be varied between the polyps (ii) there can be different calcification rates in each polyp (iii) there can be energy transfer between the polyps. Using this model we investigate the role of polyp differentiation in the morphogenesis of a coral colony,. An advantage of this approach is that coral species with radial and axial polyps (i.e. Acropora millepora) can be modeled. Potentially we can extend the morphospace of simulated morphologies significantly with the individual based approach.
References
1. J.A. Kaandorp, P.M.A. Sloot, R.M.H. Merks, R.P.M. Bak, M.J.A. Vermeij, and C. Maier Morphogenesis of the branching reef coral Madracis mirabilis, Proc. Roy. Soc. B. 272:127-133, 2005 2. Maxim V. Filatov, Jaap A. Kaandorp, Marten Postma, Robert van Liere, Kris J. Kruszy?ski, Mark J. A. Vermeij, Geert J. Streekstra, and Rolf P. M. Bak A comparison between coral colonies of the genus Madracis and simulated forms Proc. R. Soc. B published online June 23, 2010, doi:10.1098/rspb.2010.0957
Some Genomes Exhibit Explicit Periodicity in Triplet Distribution
Michael G.Sadovsky *, Eugene Yu.Bushmelev **
* Institute of computational modelling of SD RAS; **Siberian Federal university
* Correspondence: msad@icm.krasn.ru
Sadovsky pdf