Extraction of melodies in behavioural sequences

A method is proposed that extracts a set of phrases, or “melodies”, from a behavioural sequence, using a technique for extracting and compressing chains based on Information Theory. These melodies are validated by reference to a statistical criterion. An application of this method to the analysis of the behavioural sequences of two groups of mice, the first observed during the day, the second during the night, is described. The advantages and the limitations of the method are discussed.     

Humoral anti-idiotypic and anti-anti-idiotypic immune response in cancer patients treated with monoclonal antibody 17-1A

 A group of 96 patients with advanced colorectal carcinoma were treated with the mouse (m) or chimeric (c) (mouse variable regions × human IgG1 constant regions) monoclonal antibody (mAb) 17-1A recognizing the tumour-associated antigen GA733-2. Eighty-two of the 83 patients treated with mmAb17-1A and 69% of the patients given cmAb17-1A (n = 13) developed anti-idiotypic antibodies (ab₂). Auto-antibodies binding to tumour cells expressing GA733-2 were found in 7% of the patients. In a further 38 patients (40%) antitumour-cell antibodies, i.e. anti-anti-idiotypic antibodies (ab₃), were induced by the mAb17-1A therapy. Patients with detectable ab₃ after treatment had significantly higher ab₂ levels than those not developing ab₃. Addition of granulocyte/macrophage-colony-stimulating factor (GM-CSF) to mmAb17-1A significantly enhanced the induction of ab₂ as well as induction of anti-anti-idiotypic antibodies (ab₃), compared to mmAb17-1A alone. Patients with a high increase in antitumour-cell antibodies (ab₃) induced by the therapy lived significantly longer than patients with no or a low level of induction of ab₃ (P = 0.016). The results indicate that induction of an idiotypic network response might be an important effector mechanism in mAb therapy. Received: 20 October 1995 / Accepted: 18 December 1995     

Stochastic determinants of assemblage patterns in coral reef fishes: a quantification by means of two models

Synopsis One perspective emphasizing the importance of stochastic processes in determining coral reef fish assemblages implies that there is little organization in species richness, abundance structure, and spatial distribution. We examine the degree to which this perspective is correct by analyzing distribution of fishes on a collection of patch reefs (Discovery Bay, Jamaica). We ask the question whether these patches accumulate species and individuals in a manner consistent with stochastic expectations. To address this question we use two conceptual models, each permitting a different insight. One assumes that fish are distributed stochastically on patches while the other assumes presence of restrictions on fish distribution due to habitat structure. For each conceptual model we use two types of benchmark: we compare observed patterns to those predicted by theoretical models, and we also compare observed patterns to those obtained from a random reallocation of fish individuals to patches. We found that the conceptual model assuming stochastic processes appeared to provide weaker explanation of patterns than the conceptual model that includes restrictions due to habitat structure. Further, and more importantly, we found that (i) the community is shaped by a mixture of stochastic and non-stochastic mechanisms, and (ii) the stochastic assembly processes decrease in importance for species restricted to fewer microhabitat types and sites. Our study therefore indicates that patches do accumulate individuals and species in a manner consistent with stochastic expectations, however, this applies primarily to the habitat generalists (unrestricted species). By the same token, increased habitat specialization by some species imposes constraints on the stochastic model such that it eventually fails.     

Virtual resource development in the glycosciences

The development of Internet-based virtual resources is a relatively new area of scientific and technical activity that is currently undergoing rapid expansion. Major factors fuelling recent growth include the emergence of multimedia capabilities through the rapid evolution of the World Wide Web, the reduction in cost of high quality personal computers and graphics workstations and the provision of mass-marketed provider services. Prior to 1995 the presence of Internet resources in the glycosciences was virtually non-existent. Existing scientific knowledge was primarily made available on the Net through the provision of databases from gopher and ftp sites. A particular example in the glycosciences is the Carbbank database of biological carbohydrate sequences. We will describe here our efforts in 1994–95 in establishing The Glycoscience Network (TGN, http://bellatrix.pcl.ox.ac.uk/TGN/). These activities included the establishment of a newsgroup, mailing lists, Web resources and the running of the First Electronic Glycoscience Conference (EGC-1, http://bellatrix.pcl.ox.ac.uk/egc/). EGC-1 included many novel initiatives in the glycosciences including electronic posters and papers, a Virtual Conference Centre, a Web-based hyperglossary, Virtual Trade and Employment Centres, refereed electronic publishing, and the creation of a Virtual Reality Gallery. We would like to look towards the near future and discuss several initiatives in virtual resource creation that we believe will have significant scientific impact on the glycosciences including the development of bioinformatics-based servers, sophisticated interactive databases, and videoconferencing. Furthermore, we cherish the belief that these resources will foster international scientific collaboration and progress of an extent never previously possible. Finally, we indulge in speculation and make some suggestions on the form and long-term impact of Glycoscience Virtual Resources. We predict that their development may completely reconstruct the scientific environment that we work in as scientists and we reflect on the probable benefits and pitfalls to be encountered.This paper was presented at the First Electronic Glycoscience Conference (EGCI) on the World Wide Web, September 1995.     

Insights into the molecular basis of thermal stability from the structure determination of Pyrococcus furiosus gluatamate dehydrogenase

Abstract: The structure determination of the glutamate dehydrogenase from the hyperthermophile Pyrococcus furiosus has been completed at 2.2 Å resolution. The structure has been compared with the glutamate dehydrogenases from the mesophiles Clostridium symbiosum, Escherichia coli and Neurospora crassa . This comparison has revealed that the hyperthermophilic enzyme contains a striking series of networks of ion-pairs which are formed by regions of the protein which contain a high density of charged residues. Such regions are not found in the mesophilic enzymes and the number and extent of ion-pair formation is much more limited. The ion-pair networks are clustered at both inter domain and inter subunit interfaces and may well represent a major stabilising feature associated with the adaptation of enzymes to extreme temperatures.     

Six things Popper would like biologists not to ignore: In memoriam, Karl Raimund Popper, 1902–1994

To honour the memory of Sir Karl Popper, I put forward six elements of his philosophy which might be of particular interest to biologists and to philosophers of biology and which I think Popper would like them not to ignore, even if they disagree with him. They are: the primacy of problems; the criticizability of metaphysics (and thus the dubiousness of materialism); how downward causation might be real; how norms should matter to scientists; why dogmatism should be avoided; how genuine science is recognizable. I preface these six things with a brief discussion of Popper's early (but later recanted) mistakes concerning biology.     

The evolution of phage lysis timing

Summary The effect of host quantity and host quality on the evolution of phage lysis timing is analysed using marginal value theorem of optimal foraging theory. Both factors have been shown to strongly influence the latent period. A high host density selects for short latent period, which is the same result as previous investigators have found. A good host quality also promotes a short latent period. However, elasticity analysis shows that these two factors exert their influences under different sets of conditions. When host density is low, the host density is more important in determining the length of latent period, whereas when host density is high, the host quality is more important.     

Optimal flower lifetimes

Summary ESS floral lifetimes satisfy the product theorem from sex allocation theory. The dimensionless time investment per flower is a symmetric function of two dimensionless gain : cost ratios, one for each gender function.