77 research outputs found
Spatial confounding in Bayesian species distribution modeling
No full text1) Species distribution models (SDMs) are currently the main tools to derive species niche estimates and spatially explicit predictions for species geographical distribution. However, unobserved environmental conditions and ecological processes may confound the model estimates if they have direct impact on the species and, at the same time, they are correlated with the observed environmental covariates. This, so-called spatial confounding, is a general property of spatial models and it has not been studied in the context of SDMs before. 2) We examine how the estimation accuracy of SDMs depends on the type of spatial confounding. We construct two simulation studies where we alter spatial structures of the observed and unobserved covariates and the level of dependence between them. We fit generalized linear models with and without spatial random effects applying Bayesian inference and recording the bias induced to model estimates by spatial confounding. After this we examine spatial confounding also with real vegetation data from northern Norway. 3) Our results show that model estimates for coarse scale covariates, such as climate covariates, are likely to be biased if a species distribution depends also on an unobserved covariate operating on a finer spatial scale. Pushing higher probability for a relatively weak and smoothly varying spatial random effect compared to the observed covariates improved the model's estimation accuracy. The improvement was independent of the actual spatial structure of the unobserved covariate. 4) Our study addresses the major factors of spatial confounding in SDMs and provides a list of recommendations for pre-inference assessment of spatial confounding and for inference-based methods to decrease the chance of biased model estimates.Peer reviewe
Statistical studies on bacterial transmission and community dynamics : with a special emphasis on the colonization dynamics of Streptococcus pneumoniae during early childhood
No full textA central goal in science is to learn from observations about the process that generated the observations. The principles of statistical inference describe a systematic approach for such learning, in which prior information, knowledge about the underlying mechanisms and the observed data can be combined. In practice, lack of mathematical tractability, huge amounts of missing information, and the sensitivity of the conclusions on the assumptions made represent genuine challenges in the theoretically sound statistical framework. Statistical studies on the dynamics of infectious diseases easily face all these problems at once.
In the thesis we present case-studies in which the datasets on bacterial diversity, mostly on Streptococcus pneumoniae, described in terms of either genotypes or serotypic strains, are analysed. By utilizing the machinery of modern computational statistics different strategies for inference are formulated, which aim to take the special characteristics of each of the studied problem into account, while overcoming the previously mentioned challenges in computational studies. For instance, an approximate Bayesian computation scheme is formulated for analysing cross-sectional strain prevalence data and an importance sampling scheme for analysing transmission trees with a priori known complex features. The obtained results unravel the mechanisms of seasonality in pneumococcal carriage, consequences of the host population structure and the nature of within-host competition between the bacterial strains.Tilastotieteessä keskeinen päämäärä on kehitellä työkaluja, joilla aineistosta voidaan päätellä miten aineisto on syntynyt. Haasteeksi muodostuu se, että jokaisessa yksittäisessä kysymyksessä ja aineistossa on erityispiirteensä, jotka tässä päättelyssä tulisi ottaa huomioon. Eräs tapa tehdä näin on rakentaa matemaattinen malli tutkittavalle ilmiölle, tutkia kuinka hyvin malli sopii havaintoihin ja tarkastella millaiset malliparametrit vaikuttavat uskottavimmilta aineiston valossa. Ihanteena on tarkastella mahdollisimman todenmukaista mallia, koska silloin mallin parametreilla on vastineensa tosimaailmassa. Realistiset mallit ovat kuitenkin usein hankalampia analysoida, ja niiden vertaaminen havaintoihin asettaa myös aineistolle korkeat vaatimukset.
Väitöskirjassa esitetään erilaisia strategioita, joiden avulla erilaisista havainnoista voidaan päätellä, millaista bakteerien tartunta- ja populaatiodynamiikka on. Siitä miten pneumokokkibakteerin eri kantoja esiintyy yksittäisissä lapsissa, kotitalouksissa tai päiväkodeissa päätellään millaisella intensiteetillä tartuntoja tapahtuu erilaisten yksilöiden välillä, miten pitkään bakteeria kannetaan, sekä miten nämä tekijät riippuvat esimerkiksi vuodenajasta. Analyysi paljastaa myös, miten voimakasta pneumokokin lajinsisäinen kilpailu on. Lisäksi väitöskirjassa tarkastellaan, kuinka bakteerien genotyypeistä voidaan päätellä varsinaisia tartuntaketjuja, eli kuka tartutti taudin kellekin. Tämän päättelyyn ehdotetaan menetelmää, jonka puitteissa voidaan tuoda mukaan a priori tunnettuja tartuntaprosessin monimutkaisiakin piirteitä. Lisäksi tarkastellaan, miten bakteerien genotyypeistä voitaisiin nähdä millaista rakennetta isäntäpopulaatiossa on, eli millä tavalla isännät ovat toistensa kanssa tekemisissä tartuntojen näkökulmasta. Käytettyjen menetelmien filosofia pohjautuu Bayesläiseen tilastotieteeseen, jossa epävarmuuden eri lähteet pyritään ottamaan koherentisti huomioon, jolloin voidaan myös suoraan sanoa, kuinka epävarmoja esitetyt johtopäätökset ovat.ei saavutettav
Kauneus Charlotte Brontën Jane Eyre - ja Jean Rhysin Wide Sargasso Sea -teoksissa
No full textTämä pro gradu -tutkielma käsittelee kauneutta ja sen merkitystä Charlotte Brontën Jane Eyre- (julkaistu 1847) ja Jean Rhysin Wide Sargasso Sea (julkaistu 1966) -teoksissa. Jane Eyre kertoo nimikkohahmonsa tarinan ja kasvun orvosta tytöstä oppineeksi kotiopettajattareksi, joka rakastuu työnantajaansa herra Rochesteriin, jolla onkin jo vaimo, joka on suljettu mielenvikaisena kartanon ullakolle. Wide Sargasso Sea kertoo tämän vaimon tarinan, alkaen hänen lapsuudestaan ja loppuen tälle ullakolle. Teos on postkolonialistinen vastakirjoitus Jane Eyre -teokselle ja tuo esille kulttuurienvälisiä jännitteitä kolonialismin ajan Karibialla sekä antaa hullulle vaimolle ullakolla vihdoin oman äänen ja tarinan.
Tämä tutkielman tavoitteena on tarkastella sitä, millainen kauneuskäsitys teoksissa on ja miten ulkonäkö tulee teoksissa esille ja mikä sen merkitys on. Tarkastelen myös katseen käsitettä ja eri katsomisen tapoja ja seikkoja, jotka katseeseen vaikuttavat. Katseen yhteydessä tarkastelen myös pahan silmän käsitettä. Tutkin lopuksi, miten kauneus ja katse vaikuttavat teosten päähenkilöiden identiteettiin ja toimijuuteen osana laajempaa pohdintaa siitä, mikä merkitys kauneudella on laajemmin intersektionaalisessa tutkimuksessa.
Tutkimus on luonteeltaan vertailevaa. Pohjaan tutkimuksen intersektionaaliselle ja feministiselle kirjallisuudentutkimukselle ja analysoin teoksia kauneuden ja katseen teemojen kautta. Tutkimus keskittyy pääasiallisesti henkilöhahmojen analyysiin sekä kohdistuu tarkastelemaan etenkin teosten henkilöhahmokuvausta. Tutkielmassa korostuu myös teosten kulttuurillisen ja historiallisen kontekstin tarkastelu.
Totean, että 1800-luvun alkupuolen ja laajemmin viktoriaanisen ajan kauneuskäsitys vaikuttaa teosten kauneuskäsitykseen sekä laajemmin siihen, miten teosten pää- ja sivuhenkilöt katsovat maailmaa. Tämä historiallinen konteksti sisältää myös kolonialismin ja rasismin vaikutuksia tähän kauneuskäsitykseen, joka tulee molemmissa teoksissa esille eri tavoin. Koska Wide Sargasso Sea on kriittinen vastakirjoitus Jane Eyre -teokselle, se tuo näkyväksi ja kritisoi kolonialismia ja rasismia ja niiden vaikutusta siihen, miten henkilöhahmot suhtautuvat toisiinsa. Katseen tutkimisen kautta tulee ilmi, että katseella on valtaa ja katseen ja kauneuden suhde on tiivis. Analyysini kautta tulkitsen, että ulkonäöllä on suuri merkitys ja vaikutus etenkin naishenkilöhahmon identiteettiin ja toimijuuteen, ja etenkin kauneus, tai sen puute, vaikuttaa suuresti naishenkilöhahmoon. Väitän, että ulkonäkö on yksi merkittävä tekijä intersektionaalisessa tarkastelutavassa
The spread of a wild plant pathogen is driven by the road network
No full textSpatial analyses of pathogen occurrence in their natural surroundings entail unique opportunities for assessing in vivo drivers of disease epidemiology. Such studies are however confronted by the complexity of the landscape driving epidemic spread and disease persistence. Since relevant information on how the landscape influences epidemiological dynamics is rarely available, simple spatial models of spread are often used. In the current study we demonstrate both how more complex transmission pathways could be incorpoted to epidemiological analyses and how this can offer novel insights into understanding disease spread across the landscape. Our study is focused on Podosphaera plantaginis, a powdery mildew pathogen that transmits from one host plant to another by wind-dispersed spores. Its host populations often reside next to roads and thus we hypothesize that the road network influences the epidemiology of P. plantaginis. To analyse the impact of roads on the transmission dynamics, we consider a spatial dataset on the presence-absence records on the pathogen collected from a fragmented landscape of host populations. Using both mechanistic transmission modeling and statistical modeling with road-network summary statistics as predictors, we conclude the evident role of the road network in the progression of the epidemics: a phenomena which is manifested both in the enhanced transmission along the roads and in infections typically occurring at the central hub locations of the road network. We also demonstrate how the road network affects the spread of the pathogen using simulations. Jointly our results highlight how human alteration of natural landscapes may increase disease spread.</div
Approximate Bayesian computation.
No full textApproximate Bayesian computation (ABC) constitutes a class of computational methods rooted in Bayesian statistics. In all model-based statistical inference, the likelihood function is of central importance, since it expresses the probability of the observed data under a particular statistical model, and thus quantifies the support data lend to particular values of parameters and to choices among different models. For simple models, an analytical formula for the likelihood function can typically be derived. However, for more complex models, an analytical formula might be elusive or the likelihood function might be computationally very costly to evaluate. ABC methods bypass the evaluation of the likelihood function. In this way, ABC methods widen the realm of models for which statistical inference can be considered. ABC methods are mathematically well-founded, but they inevitably make assumptions and approximations whose impact needs to be carefully assessed. Furthermore, the wider application domain of ABC exacerbates the challenges of parameter estimation and model selection. ABC has rapidly gained popularity over the last years and in particular for the analysis of complex problems arising in biological sciences (e.g., in population genetics, ecology, epidemiology, and systems biology)
Species ecology can bias population estimates
No full textPopulation indices are summary statistics computed from population monitoring data and routinely used in population management and ecological studies. Changes in a population index are assumed to reflect similar changes in population size, but the exact form of this relationship is often untested. We show that non-linear population index-size relationships lead to biased population change estimates. We then analyse the reliability of eight common population indices for a wide array of different types of species, showing that both systematic bias and considerable amount of error can arise from the interaction of ecological processes and survey methods. We describe the types of scenarios (e.g. types of species and how they are monitored) in which bias occurs, and what is the likely direction of the bias in each of them. Many typical population indices are biased; for example, those measuring abundance can often overestimate change, unless species individuals tend to establish completely non-overlapping territories, and those measuring the area utilized or species presence more likely underestimate changes in population size, especially when species' individuals or population processes lead to repulsion between individuals, or when there's limited amount of suitable resources to share. By pinpointing such scenarios our results help interpretation and design of population monitoring studies and therefore can improve the reliability and efficiency of conservation and management tasks.Peer reviewe
from Strain Prevalence Data
No full textStreptococcus pneumoniae is a typical commensal bacterium causing severe diseases. Its prevalence is high among young children attending day care units, due to lower levels of acquired immunity and a high rate of infectious contacts between the attendees. Understanding the population dynamics of different strains of S.pneumoniae is necessary, for example, for making successful predictions of changes in the composition of the strain community under intervention policies. Here we analyze data on the strains of S. pneumoniae carried in attendees of day care units in the metropolitan area of Oslo, Norway. We introduce a variant of approximate Bayesian computation methods, which is suitable for estimating the parameters governing the transmission dynamics in a setting where small local populations of hosts are subject to epidemics of different pathogenic strains due to infections independently acquired from the community. We find evidence for strong between‐strain competition, as the acquisition of other strains in the already colonized hosts is estimated to have a relative rate of 0.09 (95% credibility interval [0.06, 0.14]). We also predict the frequency and size distributions for epidemics within the day care unit, as well as other epidemiologically relevant features. The assumption of ecological neutrality between the strains is observed to be compatible with the data. Model validation checks and the consistency of our results with previous research support the validity of our conclusions
Posterior medians and 95% credibility intervals for the hyperparameters of the statistical model, where <i>ρ</i> describes the 1st order autocorrelation, nominal variance and range describes the spatial random field, where nominal variance describes the overall variance of the field and range corresponds to the distance after which the spatial autocorrelation is estimated to become smaller than 0.1, when the Matern covariance structure is assumed.
No full textPosterior medians and 95% credibility intervals for the hyperparameters of the statistical model, where ρ describes the 1st order autocorrelation, nominal variance and range describes the spatial random field, where nominal variance describes the overall variance of the field and range corresponds to the distance after which the spatial autocorrelation is estimated to become smaller than 0.1, when the Matern covariance structure is assumed.</p
The spread of a wild plant pathogen is driven by the road network - Fig 4
No full textThe two computed centrality measures, betweenness (A) and closeness (B), for the considered host populations, computed based on their projection to the closest point in the road network. The correlation between the Euclidean- and shortest distance by road for a random set of pairs of host populations (C) and the relationship between the computed betweenness summary-statistic and the presence and absence of pathogen in different years (D). The roadmaps in the background were created using data produced by National Land Survey of Finland.</p
- …
