Solve Sæbø

Kin-related spatial structure in brown bears Ursus arctos

Kin-related social structure may influence reproductive success and survival and, hence, the dynamics of populations. It has been documented in many gregarious animal populations, but few solitary species. Using molecular methods and field data we tested: (1) whether kin-related spatial structure exists in the brown bear (Ursus arctos), which is a solitary carnivore, (2) whether home ranges of adult female kin overlap more than those of nonkin, and (3) whether multigenerational matrilinear assemblages, i.e., aggregated related females, are formed. Pairwise genetic relatedness between adult (5 years and older) female dyads declined significantly with geographic distance, whereas this was not the case for male–male dyads or opposite sex dyads. The amount of overlap of multiannual home ranges was positively associated with relatedness among adult females. This structure within matrilines is probably due to kin recognition. Plotting of multiannual home-range centers of adult females revealed formation of two types of matrilines, matrilinear assemblages exclusively using an area and dispersed matrilines spread over larger geographic areas. The variation in matrilinear structure might be due to differences in competitive abilities among females and habitat limitations. The influence of kin-related spatial structure on inclusive fitness needs to be clarified in solitary mammals.

Should I stay or should I go? Natal dispersal in the brown bear

We studied the causes of natal dispersal of male and female brown bears, Ursus arctos, in two study areas in Sweden. Males had a higher dispersal probability (94%) than females (41%). For males, we found no difference in dispersal probability or mean age of dispersal between the study areas, in spite of differences in population density and sex ratio. Male–male competition did not seem to influence subadult male dispersal probability significantly. These results support the inbreeding avoidance hypothesis as the cause of male natal dispersal. For females, dispersal probability decreased with increasing maternal age and decreased with increasing body size, and an interaction between maternal age and body size suggested that the importance of body size decreased with increasing maternal age. Nondispersing females were closer to their mother than their dispersing sibling sisters were in the period between weaning and dispersal. Female littermates seemed to compete for philopatry, suggesting that a dominance hierarchy among female littermates based on body size may cause the subdominant sister to disperse. If juvenile females are born into matrilineal assemblages, surrounded mostly by related females, the competition for philopatry may not be as severe as when they are born into an area surrounded by mostly nonkin females. This hypothesis is supported by the decreasing importance of body size for dispersal with increasing maternal age. We suggest that natal dispersal in juvenile female brown bears can be explained by the resident fitness hypothesis.

Gene expression profiling of peripheral blood cells for early detection of breast cancer

IntroductionEarly detection of breast cancer is key to successful treatment and patient survival. We have previously reported the potential use of gene expression profiling of peripheral blood cells for early detection of breast cancer. The aim of the present study was to refine these findings using a larger sample size and a commercially available microarray platform.MethodsBlood samples were collected from 121 females referred for diagnostic mammography following an initial suspicious screening mammogram. Diagnostic work-up revealed that 67 of these women had breast cancer while 54 had no malignant disease. Additionally, nine samples from six healthy female controls were included. Gene expression analyses were conducted using high density oligonucleotide microarrays. Partial Least Squares Regression (PLSR) was used for model building while a leave-one-out (LOO) double cross validation approach was used to identify predictors and estimate their prediction efficiency.ResultsA set of 738 probes that discriminated breast cancer and non-breast cancer samples was identified. By cross validation we achieved an estimated prediction accuracy of 79.5% with a sensitivity of 80.6% and a specificity of 78.3%. The genes deregulated in blood of breast cancer patients are related to functional processes such as defense response, translation, and various metabolic processes, such as lipid- and steroid metabolism.ConclusionsWe have identified a gene signature in whole blood that classifies breast cancer patients and healthy women with good accuracy supporting our previous findings.

A Gene Expression Pattern in Blood for the Early Detection of Alzheimer's Disease

A whole genome screen was performed using oligonucleotide microarray analysis on blood from a large clinical cohort of Alzheimers disease (AD) patients and control subjects as clinical sample. Blood samples for total RNA extraction were collected in PAXgene tubes, and gene expression analysis performed on the AB1700 Whole Genome Survey Microarrays. When comparing the gene expression of 94 AD patients and 94 cognitive healthy controls, a Jackknife gene selection based method and Partial Least Square Regression (PLSR) was used to develop a disease classifier algorithm, which gives a test score indicating the presence (positive) or absence (negative) of AD. This algorithm, based on 1239 probes, was validated in an independent test set of 63 subjects comprising 31 AD patients, 25 age-matched cognitively healthy controls, and 7 young controls. This algorithm correctly predicted the class of 55/63 (accuracy 87%), including 26/31 AD samples (sensitivity 84%) and 29/32 controls (specificity 91%). The positive likelihood ratio was 8.9 and the area under the receiver operating characteristic curve (ROC AUC) was 0.94. Furthermore, the algorithm also discriminated AD from Parkinsons disease in 24/27 patients (accuracy 89%). We have identified and validated a gene expression signature in blood that classifies AD patients and cognitively healthy controls with high accuracy and show that alterations specific for AD can be detected distant from the primary site of the disease.

A Partial Least Squares based algorithm for parsimonious variable selection

BackgroundIn genomics, a commonly encountered problem is to extract a subset of variables out of a large set of explanatory variables associated with one or several quantitative or qualitative response variables. An example is to identify associations between codon-usage and phylogeny based definitions of taxonomic groups at different taxonomic levels. Maximum understandability with the smallest number of selected variables, consistency of the selected variables, as well as variation of model performance on test data, are issues to be addressed for such problems.ResultsWe present an algorithm balancing the parsimony and the predictive performance of a model. The algorithm is based on variable selection using reduced-rank Partial Least Squares with a regularized elimination. Allowing a marginal decrease in model performance results in a substantial decrease in the number of selected variables. This significantly improves the understandability of the model. Within the approach we have tested and compared three different criteria commonly used in the Partial Least Square modeling paradigm for variable selection; loading weights, regression coefficients and variable importance on projections. The algorithm is applied to a problem of identifying codon variations discriminating different bacterial taxa, which is of particular interest in classifying metagenomics samples. The results are compared with a classical forward selection algorithm, the much used Lasso algorithm as well as Soft-threshold Partial Least Squares variable selection.ConclusionsA regularized elimination algorithm based on Partial Least Squares produces results that increase understandability and consistency and reduces the classification error on test data compared to standard approaches.

Lasting behavioural responses of brown bears to experimental encounters with humans

Summary 1. Some large carnivore populations are increasing in Europe and North America, and minimizing interactions between people and carnivores is a major management task. Analysing the effects of human disturbance on wildlife from a predator–prey perspective is also of conservation interest, because individual behavioural responses to the perceived risk of predation may ultimately influence population distribution and demography. 2. The Scandinavian brown bear population provides a good model to study the interactions between an expanding large carnivore population, and people who use forests extensively for professional and recreational activities. We experimentally approached 52 GPS-collared brown bears (293 approaches on foot) from 2006 to 2011, to document the reaction of bears and quantify the effect of disturbance on bear movements. 3. None of the bears reacted aggressively to the observers. Although the location of the animals was known, bears were usually in quite concealed spots and were physically detected in only 16% of the approaches (seen in 42 approaches; heard in 6). However, the bears altered their daily movement patterns after the approaches. Bears increased movement at night-time and moved less at daytime, which was most visible in days 1 and 2 after the approaches, altering their foraging and resting routines. 4. Synthesis and applications. We provide experimental evidence on the effect of human disturbance on a large carnivore. The lack of aggressive reactions to approaching observers reinforces the idea that European brown bears generally avoid people, although bears can respond aggressively if they feel threatened (e.g. when wounded). However, the movement patterns of the bears changed after disturbance. Separating large carnivores and people temporally and spatially is an important goal for conservation and management. Conserving the shrub cover that provides concealment to the carnivores and keeping people away from the most densely vegetated spots in the forests is a way to avoid encounters between carnivores and people, therefore promoting human safety and carnivore conservation.

Light traps fail to estimate reliable malaria mosquito biting rates on Bioko Island, Equatorial Guinea

BackgroundThe human biting rate (HBR), an important parameter for assessing malaria transmission and evaluating vector control interventions, is commonly estimated by human landing collections (HLC). Although intense efforts have been made to find alternative non-exposure mosquito collection methods, HLC remains the standard for providing reliable and consistent HBRs. The aim of this study was to assess the relationship between human landing and light trap collections (LTC), in an attempt to estimate operationally feasible conversion factors between the two. The study was conducted as part of the operational research component of the Bioko Island Malaria Control Project (BIMCP), Equatorial Guinea.MethodsMalaria mosquitoes were collected indoors and outdoors by HLCs and LTCs in three villages on Bioko Island, Equatorial Guinea during five bimonthly collections in 2009. Indoor light traps were suspended adjacent to occupied long-lasting, insecticide-treated bed nets. Outdoor light traps were placed close to the outer wall under the roof of the collection house. Collected specimens were subjected to DNA extraction and diagnostic PCR to identify species within the Anopheles gambiae complex. Data were analysed by simple regression of log-transformed values and by Bayesian regression analysis.ResultsThere was a poor correlation between the two collection methods. Results varied by location, venue, month, house, but also by the statistical method used. The more robust Bayesian analyses indicated non-linear relationships and relative sampling efficiencies being density dependent for the indoor collections, implying that straight-forward and simple conversion factors could not be calculated for any of the locations. Outdoor LTC:HLC relationships were weak, but could be estimated at 0.10 and 0.07 for each of two locations.ConclusionsLight trap collections in combination with bed nets are not recommended as a reliable method to assess human biting rates on Bioko Island. Different statistical analyses methods give variable and inconsistent results. Substantial variation in collection methods prevents the determination of reliable and operationally feasible conversion factors for both indoor and outdoor data. Until improved mosquito collection methods are developed that can provide reliable and unbiased HBR estimates, HLCs should continue to serve as the reference method for HBR estimation.

Exact computation of the distribution of likelihood ratios with forensic applications.

If complex DNA profiles, conditioned on multiple individuals are evaluated, it may be difficult to assess the strength of the evidence based on the likelihood ratio. A likelihood ratio does not give information about the relative weights that are provided by separate contributors. Alternatively, the observed likelihood ratio can be evaluated with respect to the distribution of the likelihood ratio under the defense hypothesis. We present an efficient algorithm to compute an exact distribution of likelihood ratios that can be applied to any LR-based model. The distribution may have several applications, but is used here to compute a p-value that corresponds to the observed likelihood ratio. The p-value is the probability that a profile under the defense hypothesis, substituted for a questioned contributor e.g. suspect, would attain a likelihood ratio which is at least the same magnitude as that observed. The p-value can be thought of as a scaled version of the likelihood ratio, giving a quantitative measure of the strength of the evidence relative to the specified hypotheses and the model used for the analysis. The algorithm is demonstrated on examples based on real data. R code for the algorithm is freely available in the R package euroMix.

Mining for genotype-phenotype relations in Saccharomyces using partial least squares

BackgroundMultivariate approaches are important due to their versatility and applications in many fields as it provides decisive advantages over univariate analysis in many ways. Genome wide association studies are rapidly emerging, but approaches in hand pay less attention to multivariate relation between genotype and phenotype. We introduce a methodology based on a BLAST approach for extracting information from genomic sequences and Soft- Thresholding Partial Least Squares (ST-PLS) for mapping genotype-phenotype relations.ResultsApplying this methodology to an extensive data set for the model yeast Saccharomyces cerevisiae, we found that the relationship between genotype-phenotype involves surprisingly few genes in the sense that an overwhelmingly large fraction of the phenotypic variation can be explained by variation in less than 1% of the full gene reference set containing 5791 genes. These phenotype influencing genes were evolving 20% faster than non-influential genes and were unevenly distributed over cellular functions, with strong enrichments in functions such as cellular respiration and transposition. These genes were also enriched with known paralogs, stop codon variations and copy number variations, suggesting that such molecular adjustments have had a disproportionate influence on Saccharomyces yeasts recent adaptation to environmental changes in its ecological niche.ConclusionsBLAST and PLS based multivariate approach derived results that adhere to the known yeast phylogeny and gene ontology and thus verify that the methodology extracts a set of fast evolving genes that capture the phylogeny of the yeast strains. The approach is worth pursuing, and future investigations should be made to improve the computations of genotype signals as well as variable selection procedure within the PLS framework.

Multi-level binary replacement (MBR) design for computer experiments in high-dimensional nonlinear systems

Computer experiments are useful for studying a complex system, e.g. a high‐dimensional nonlinear mathematical model of a biological or physical system. Based on the simulation results, an empirical “metamodel” may then be developed, emulating the behavior of the model in a way that is faster to compute and easier to understand. In modelometrics, the model phenome of a computer model is recorded, once and for all, by structured simulations according to a factorial design in the model inputs, and with high‐dimensional profiling of its simulation outputs. A multivariate metamodel is then developed, by multivariate analysis of the input–output data, akin to how high‐dimensional data are analyzed in chemometrics. To reveal strongly nonlinear input–output relationships, the factorial design must probe the design space at many different levels for each of the many input factors. A reduced factorial design method may be required if combinatorial explosion is to be avoided. In the multi‐level binary replacement (MBR) design the levels of each input factor are represented as binary numbers, and all the individual binary factor bits are then combined in a fractional factorial (FF) design. The experiment size can thereby be greatly reduced at the price of some binary confounding. The MBR method is here described and then illustrated for the optimization of a nonlinear model of a microbiological growth curve with five design factors, for finding the relevant region in the design space, and subsequently for estimating the optimal design points in that space. Copyright