A censored quantile regression approach for relative survival analysis: Relative survival quantile regression

We propose a censored quantile regression model for the analysis of relative survival data. We create a hybrid data set consisting of the study observations and counterpart randomly sampled pseudopopulation observations imputed from population life tables that adjust for expected mortality. We then fit a censored quantile regression model to the hybrid data incorporating demographic variables (e.g., age, biologic sex, calendar time) corresponding to the population life tables of demographically-similar individuals, a population versus study covariate, and its interactions with the variables of interest. These latter variables can be interpreted as relative survival parameters that depict the differences in failure quantiles between the study participants and their population counterparts.     

Smoothed quantile regression analysis of competing risks

Censored quantile regression models, which offer great flexibility in assessing covariate effects on event times, have attracted considerable research interest. In this study, we consider flexible estimation and inference procedures for competing risks quantile regression, which not only provides meaningful interpretations by using cumulative incidence quantiles but also extends the conventional accelerated failure time model by relaxing some of the stringent model assumptions, such as global linearity and unconditional independence. Current method for censored quantile regressions often involves the minimization of the L₁‐type convex function or solving the nonsmoothed estimating equations. This approach could lead to multiple roots in practical settings, particularly with multiple covariates. Moreover, variance estimation involves an unknown error distribution and most methods rely on computationally intensive resampling techniques such as bootstrapping. We consider the induced smoothing procedure for censored quantile regressions to the competing risks setting. The proposed procedure permits the fast and accurate computation of quantile regression parameter estimates and standard variances by using conventional numerical methods such as the Newton–Raphson algorithm. Numerical studies show that the proposed estimators perform well and the resulting inference is reliable in practical settings. The method is finally applied to data from a soft tissue sarcoma study.     

Flow cytometric analysis of nuclear DNA content in Musa

 Nuclear genome size variation was studied in Musa acuminata (A genome), Musa balbisiana (B genome) and a range of triploid clones differing in genomic constitution (i.e. the relative number of A and B genomes). Nuclear DNA content was estimated by flow cytometry of nuclei stained by propidium iodide. The A and B genomes of Musa differ in size, the B genome being smaller by 12% on average. No variation in genome size was found among the accessions of M. balbisiana (average genome size 537 Mbp). Small, but statistically significant, variation was found among the subspecies and clones of M. acuminata (ranging from 591 to 615 Mbp). This difference may relate to the geographical origin of the individual accessions. Larger variation in genome size (8.8%) was found among the triploid Musa accessions (ranging from 559 to 613 Mbp). This variation may be due to different genomic constitutions as well as to differences in the size of their A genomes. It is proposed that a comparative analysis of genome size in diploids and triploids may be helpful in identifying putative diploid progenitors of cultivated triploid Musa clones. Statistical analysis of data on genome size resulted in a grouping which agreed fairly well with the generally accepted taxonomic classification of Musa. Received: 11 May 1998 / Accepted: 29 September 1998     

Multikink quantile regression for longitudinal data with application to progesterone data analysis

Motivated by investigating the relationship between progesterone and the days in a menstrual cycle in a longitudinal study, we propose a multikink quantile regression model for longitudinal data analysis. It relaxes the linearity condition and assumes different regression forms in different regions of the domain of the threshold covariate. In this paper, we first propose a multikink quantile regression for longitudinal data. Two estimation procedures are proposed to estimate the regression coefficients and the kink points locations: one is a computationally efficient profile estimator under the working independence framework while the other one considers the within-subject correlations by using the unbiased generalized estimation equation approach. The selection consistency of the number of kink points and the asymptotic normality of two proposed estimators are established. Second, we construct a rank score test based on partial subgradients for the existence of the kink effect in longitudinal studies. Both the null distribution and the local alternative distribution of the test statistic have been derived. Simulation studies show that the proposed methods have excellent finite sample performance. In the application to the longitudinal progesterone data, we identify two kink points in the progesterone curves over different quantiles and observe that the progesterone level remains stable before the day of ovulation, then increases quickly in 5 to 6 days after ovulation and then changes to stable again or drops slightly.     

Determination of ploidy level and nuclear DNA content in blueberry by flow cytometry

The technique of DNA flow cytometry was used to study variation in DNA content among different ploidy levels, as well as among diploid species, of Vaccinium section Cyanococcus. In a sample of plants of varying ploidy level, the relative fluorescence intensity (RFI) of nuclei stained with propidium iodide was a function of the number of chromosome sets (x), as represented by the linear equation RFI=3.7x-2.3 (r²=95%). The data indicated that DNA flow cytometry could be useful for the determination of ploidy level at the seedling stage in blueberry. They also suggest that conventional polyploid evolution has occurred in this section of the genus Vaccinium with an increase in nuclear DNA content concurrent with the increase in chromosome number. The nuclear DNA content of diploid species of Vaccinium section Cyanococcus was estimated from the relationship of the observed RFI to an internal known DNA standard (trout red blood cells). A nested analysis of variance indicated significant variation among species, as well as among populations within species, in nuclear DNA content, although this variation was small compared to the variation among ploidy levels. The variation in nuclear DNA content corresponded to the phylogenetic relationships among species determined from previous studies.     

Relationships between nuclear DNA content and seed and leaf size in soybean

 A correlation between genome size and agronomically important traits has been observed in many plant species. The goal of the present research was to determine the relationship between genome size, seed size, and leaf width and length in soybean [Glycine max (L.) Merr.] Twelve soybean strains, representing three distinct seed size groups, were analyzed. Flow cytometry was used to estimate their 2C nuclear DNA contents. Data on seed size and leaf size of the 12 strains were obtained from 1994 and 1995 field experiments. Variation of 2C nuclear DNA among the 12 soybean strains was 4.6%, ranging from 2.37 pg for a small-seed strain to 2.48 pg for a large-seed strain. Strain seed size was positively associated with leaf width (r=0.92) and leaf length (r=0.93). Genome size was highly correlated with seed size (r=0.97), leaf width (r=0.90) , and leaf length (r=0.93). The results of our study indicate that there is a significant correlation between genome size and leaf and seed size in soybean. It is possible that selection for greater seed size either leads to, or results from, greater genome size. If so, this relationship might be worth exploring at a more fundamental level. Received: 5 April 1997 / Accepted: 9 January 1998     

Smoothed quantile regression for partially functional linear models in high dimensions

Practitioners of current data analysis are regularly confronted with the situation where the heavy-tailed skewed response is related to both multiple functional predictors and high-dimensional scalar covariates. We propose a new class of partially functional penalized convolution-type smoothed quantile regression to characterize the conditional quantile level between a scalar response and predictors of both functional and scalar types. The new approach overcomes the lack of smoothness and severe convexity of the standard quantile empirical loss, considerably improving the computing efficiency of partially functional quantile regression. We investigate a folded concave penalized estimator for simultaneous variable selection and estimation by the modified local adaptive majorize-minimization (LAMM) algorithm. The functional predictors can be dense or sparse and are approximated by the principal component basis. Under mild conditions, the consistency and oracle properties of the resulting estimators are established. Simulation studies demonstrate a competitive performance against the partially functional standard penalized quantile regression. A real application using Alzheimer's Disease Neuroimaging Initiative data is utilized to illustrate the practicality of the proposed model.     

应用混合模型和分位数回归分析东黄海小黄鱼肥满度空间异质性

基于2008-2010年黄海南部近海(SYS)、东海北部外海(NECS)和东海中部近海(MECS)小黄鱼体长和体质量数据,采用均值回归和分位数回归模型,解析了小黄鱼幼鱼和成鱼群体体长-体质量关系的空间变异.结果表明: 协方差模型和线性混合模型的残差标准误基本一致,线性模型残差标准误最高.从线性混合模型对特定区域和总体区域平均体质量计算的相对比值来看,SYS和NECS幼鱼群体的平均体质量高于总体平均值,但MECS低于总体平均值;成鱼群体则为NECS平均体质量高于总体平均值,MECS和SYS低于总体平均值.分位回归估计的肥满度和异速生长指数结果显示,幼鱼群体在不同分位的估计参数呈显著变化,SYS异速生长指数均值为2.85,在0.1~0.95分位的估计值变化范围为2.63~2.96.MECS和NECS参数估计值和置信区间在各分位数呈异质性变化,低分位时,NECS估计值在3个调查区域中最低,MECS最高;高分位时,MECS和NECS均高于SYS.对低分位0.25、中分位0.5和高分位0.75分位数的异速体长体质量关系的方差分析结果显示,低分位和高分位数之间体长 体质量关系极为显著(0.25∶0.75,F=6.38,df=1737,P<0.01),低分位数和中分位数之间为显著(0.25∶0.5,F=2.35,df=1737,P=0.039),中分位数和高分位数之间接近显著(0.5∶0.75,F=2.21,df=1737,P=0.051).成鱼群体SYS异速生长指数均值为3.01,在0.1~0.95分位的估计值变化范围为2.77~3.10.低分位和高分位数之间体长 体质量关系差异达到显著水平(0.25∶0.75,F=3.31,df=2793,P=0.01),低分位和中分位之间差异不显著(0.25∶0.5,F=0.98,df=2793,P=0.43),而高分位和中分位之间则差异极显著(0.5∶0.75,F=3.56,df=2793,P<0.01).     

Quantification of nuclear DNA and G-C content in marine macroalgae by flow cytometry of isolated nuclei

Summary The amounts of nuclear DNA in ten species of seaweeds belonging to the Rhodophyceae, Phaeophyceae, and Chlorophyceae were determined by flow cytometric analysis of nuclei isolated from protoplasts. Genome size was determined from the fluorescence of the nuclei stained with ethidium bromide. The size of the nuclear genome ranged from 0.13 pg per cell in the 1 C population ofUlva rigida to 3.40 pg per cell in the 2 C population ofSphacelaria sp. GC% analysis was based on staining with either Hoechst 33342 or mithramycin A, two fluorochromes specific for the bases A-T and G-C, respectively. Two models were used for the estimation of the proportion of guanine plus cytosine in the nuclear genome. The first one was based on the linear relationships mithramycin A fluorescence/G-C content and ethidium bromide fluorescence/total DNA content. The second model, based on the curvilinear relationships Hoechst 33342 fluorescence/A-T content and mithramycin A fluorescence/G-C content, resulted in comparatively more homogenous and consistent data and appears more accurate. Comparison with previous reports from other methods for the physical investigation of nuclear genomes shows that flow cytometry of nuclei isolated from protoplasts is an accurate, convenient and robust technique to assay for genome sizes and base pair composition in marine macroalgae.Abbreviations A-T nucleic bases adenine and thymine - CRBC chicken red blood cell - FALS forward-angle light scatter - G-C nucleic bases guanine and cytosine - SEIM sorbitol enzymatic incubation medium - SWIM sea water incubation medium - Tm thermal denaturation temperature of DNA     

First nuclear DNA amounts in more than 300 angiosperms

BACKGROUND AND AIMS: Genome size (DNA C-value) data are key biodiversity characters of fundamental significance used in a wide variety of biological fields. Since 1976, Bennett and colleagues have made scattered published and unpublished genome size data more widely accessible by assembling them into user-friendly compilations. Initially these were published as hard copy lists, but since 1997 they have also been made available electronically (see the Plant DNA C-values database http://www.kew.org/cval/homepage.html). Nevertheless, at the Second Plant Genome Size Meeting in 2003, Bennett noted that as many as 1000 DNA C-value estimates were still unpublished and hence unavailable. Scientists were strongly encouraged to communicate such unpublished data. The present work combines the databasing experience of the Kew-based authors with the unpublished C-values produced by Zonneveld to make a large body of valuable genome size data available to the scientific community. METHODS: C-values for angiosperm species, selected primarily for their horticultural interest, were estimated by flow cytometry using the fluorochrome propidium iodide. The data were compiled into a table whose form is similar to previously published lists of DNA amounts by Bennett and colleagues. KEY RESULTS AND CONCLUSIONS: The present work contains C-values for 411 taxa including first values for 308 species not listed previously by Bennett and colleagues. Based on a recent estimate of the global published output of angiosperm DNA C-value data (i.e. 200 first C-value estimates per annum) the present work equals 1.5 years of average global published output; and constitutes over 12 % of the latest 5-year global target set by the Second Plant Genome Size Workshop (see http://www.kew.org/cval/workshopreport.html). Hopefully, the present example will encourage others to unveil further valuable data which otherwise may lie forever unpublished and unavailable for comparative analyses.     

The relationship between nuclear DNA content and leaf strategy in seed plants

BACKGROUND AND AIMS: Species' 2C-values (mass of DNA in G(1) phase 2n nuclei) vary by at least four orders of magnitude among seed plants. The 2C-value has been shown to be co-ordinated with a number of other species traits, and with environmental variables. A prediction that species 2C-values are negatively related to leaf life span (LL) and leaf mass per area (LMA) is tested. These leaf traits are components of a major dimension of ecological variation among plant species. METHODS: Flow cytometry was used to measure the 2C-values for 41 Australian seed plant species, 40 of which were new to the literature. Where possible, LL and LMA data from the global literature were combined with 2C-values from our data set and online C-value databases. KEY RESULTS: Across all species, weak positive relationships were found between 2C-values and both LL and LMA; however, these did not reflect the relationships within either angiosperms or gymnosperms. Across 59 angiosperm species, there were weak negative relationships between 2C-values and both LL (r2 = 0.13, P = 0.005) and LMA (r2 = 0.15, P = 0.002). These relationships were the result of shifts to longer LL and greater LMA in woody compared with herbaceous growth forms, with no relationships present within growth forms. It was not possible to explain a positive relationship between 2C-values and LMA (r2 = 0.30, P = 0.024) across 17 gymnosperm species. The 2C-value was not related to LL or LMA either across species within orders (except for LMA among Pinales), or as radiation divergences in a model phylogeny. CONCLUSIONS: Gymnosperms appear to vary along a spectrum different from angiosperms. Among angiosperms, weak negative cross-species relationships were associated with growth form differences, and traced to a few divergences deep in the model phylogeny. These results suggest that among angiosperms, nuclear DNA content and leaf strategy are unrelated.     

Evolution of genome size in fishes: a phylogenetic test of the Hinegardner and Rosen hypothesis

Despite remarkable advances in genomic studies over the past few decades, surprisingly little is known about the processes governing genome evolution at macroevolutionary timescales. In a seminal paper, Hinegardner and Rosen (Am Nat 106:621–644, 1972) suggested that taxa characterized by larger genomes should also display disproportionately stronger fluctuations in genome size. Therefore, according to the Hinegardner and Rosen (HR) hypothesis, there should be a negative correlation between average within-family genome size and its corresponding coefficient of variation (CV), a prediction that was supported by their analysis of the genomes of 275 species of fish. In this study we reevaluate the HR hypothesis using an expanded dataset (2050 genome size records). Moreover, in addition to the use of standard linear regression techniques, we also conducted modern comparative analyses that take into account phylogenetic non-independence. Our analyses failed to confirm the negative relationship detected in the original study, suggesting that the evolution of genome size in fishes might be more complex than envisioned by the HR hypothesis. Interestingly, the frequency distribution of fish genome sizes was strongly skewed, even on a logarithmic scale, suggesting that the dynamics underlying genome size evolution are driven by multiplicative phenomena, which might include chromosomal rearrangements and the expansion of transposable elements.     

New reports of nuclear DNA content for 407 vascular plant taxa from the United States

Background and Aims

The amount of DNA in an unreplicated haploid nuclear genome (C-value) ranges over several orders of magnitude among plant species and represents a key metric for comparing plant genomes. To extend previously published datasets on plant nuclear content and to characterize the DNA content of many species present in one region of North America, flow cytometry was used to estimate C-values of woody and herbaceous species collected in Wisconsin and the Upper Peninsula of Michigan, USA.

Methods

A total of 674 samples and vouchers were collected from locations across Wisconsin and Michigan, USA. From these, C-value estimates were obtained for 514 species, subspecies and varieties of vascular plants. Nuclei were extracted from samples of these species in one of two buffers, stained with the fluorochrome propidium iodide, and an Accuri C-6 flow cytometer was used to measure fluorescence peaks relative to those of an internal standard. Replicate extractions, coefficients of variation and comparisons to published C-values in the same and related species were used to confirm the accuracy and reliability of our results.

Key Results and Conclusions

Prime C-values for 407 taxa are provided for which no published data exist, including 390 angiosperms, two gymnosperms, ten monilophytes and five lycophytes. Non-prime reports for 107 additional taxa are also provided. The prime values represent new reports for 129 genera and five families (of 303 genera and 97 families sampled). New family C-value maxima or minima are reported for Betulaceae, Ericaceae, Ranunculaceae and Sapindaceae. These data provide the basis for phylogenetic analyses of C-value variation and future analyses of how C-values covary with other functional traits.     

Evolution of DNA amounts across land plants (embryophyta)

BACKGROUND AND AIMS: DNA C-values in land plants (comprising bryophytes, lycophytes, monilophytes, gymnosperms and angiosperms) vary approximately 1000-fold from approx. 0.11 to 127.4 pg. To understand the evolutionary significance of this huge variation it is essential to evaluate the phylogenetic component. Recent increases in C-value data (e.g. Plant DNA C-values database; release 2.0, January 2003; http://www.rbgkew.org.uk/cval/homepage.html) together with improved consensus of relationships between and within land plant groups makes such an analysis timely. METHODS: Insights into the distribution of C-values in each group of land plants were gained by superimposing available C-value data (4119 angiosperms, 181 gymnosperms, 63 monilophytes, 4 lycophytes and 171 bryophytes) onto phylogenetic trees. To enable ancestral C-values to be reconstructed for clades within land plants, character-state mapping with parsimony and MacClade was also applied. KEY RESULTS AND CONCLUSIONS: Different land plant groups are characterized by different C-value profiles, distribution of C-values and ancestral C-values. For example, the large ( approximately 1000-fold) range yet strongly skewed distribution of C-values in angiosperms contrasts with the very narrow 12-fold range in bryophytes. Further, character-state mapping showed that the ancestral genome sizes of both angiosperms and bryophytes were reconstructed as very small (i.e. < or =1.4 pg) whereas gymnosperms and most branches of monilophytes were reconstructed with intermediate C-values (i.e. >3.5, <14.0 pg). More in-depth analyses provided evidence for several independent increases and decreases in C-values; for example, decreases in Gnetaceae (Gymnosperms) and heterosperous water ferns (monilophytes); increases in Santalales and some monocots (both angiosperms), Pinaceae, Sciadopityaceae and Cephalotaxaceae (Gymnosperms) and possibly in the Psilotaceae + Ophioglossaceae clade (monilophytes). Thus, in agreement with several focused studies within angiosperm families and genera showing that C-values may both increase and decrease, it is apparent that this dynamic pattern of genome size evolution is repeated on a broad scale across land plants.     

Amounts of nuclear DNA in marine halophytes

The amount of nuclear DNA, expressed as the C-value, was estimated for 13 marine halophytic plant species from six families. Plant material was collected in the nature reserve of the Strunjan saltpan in the Northern Adriatic and comprised all halophytic species inside the investigated area. Reproductive region of the shoot or root tips of halophytes were dissected, nuclei were Feulgen stained and 2C-values were measured by DNA image cytometry as follows: Crithmum maritimum (4.38 pg DNA), Artemisia caerulescens (6.43 pg), Aster tripolium (21.43 pg), Inula crithmoides (3.63 pg), Atriplex portulacoides (1.83 pg), A. prostrata (1.51 pg), Salicornia europaea (2.75 pg), Salsola soda (2.62 pg), Sarcocornia fruticosa (5.91 pg), Suaeda maritima (2.11 pg), Limonium angustifolium (5.06 pg), Puccinellia palustris (8.15 pg) and Ruppia cirrhosa (4.65 pg). With the exception of the C-value estimate for A. caerulescens, which has been listed in the Plant DNA C-values Database, the C-values represent the first estimates for all the examined species. In addition, the C-value for R. cirrhosa is also the first report for the family Ruppiaceae. The investigated halophytes had a smaller genome size compared to other known C-values for species within a particular family and also when compared to the mean values of dicots and monocots. The study also showed that halophylic annuals have a smaller genome size (2.49 pg) than perennial ones (7.45 pg DNA).     

Variation across amphibian species in the size of the nuclear genome supports a pluralistic, hierarchical approach to the C-value enigma

Amphibians have featured prominently in discussions of the C-value enigma, the still-unresolved puzzle regarding the evolution of genome size. Their wide range in nuclear DNA contents and diverse ecological and developmental lifestyles make them excellent subjects for addressing the key elements of the C-value enigma. However, in some cases the importance of work on amphibians appears to be overstated. This is especially true of claims that patterns of variation in salamanders support a particular theory of genome size evolution to the exclusion of others. This study provides a critical re-examination of some of these claims, as well as an investigation of the relationships between genome size, cell and nuclear size, and metabolism in amphibians. The results of these analyses, combined with an overview of previous amphibian genome size literature, strongly indicate the need for a pluralistic approach to the C-value enigma. In particular, it must be recognized that evolutionary forces operating and interacting at several levels of biological organization (of which the genome itself is one) are responsible for the observed patterns in amphibian genome size distributions.  © 2003 The Linnean Society of London, Biological Journal of the Linnean Society , 2003, 79 , 329–339.     

Patterns of maximum body size evolution in Cenozoic land mammals: eco-evolutionary processes and abiotic forcing

There is accumulating evidence that macroevolutionary patterns of mammal evolution during the Cenozoic follow similar trajectories on different continents. This would suggest that such patterns are strongly determined by global abiotic factors, such as climate, or by basic eco-evolutionary processes such as filling of niches by specialization. The similarity of pattern would be expected to extend to the history of individual clades. Here, we investigate the temporal distribution of maximum size observed within individual orders globally and on separate continents. While the maximum size of individual orders of large land mammals show differences and comprise several families, the times at which orders reach their maximum size over time show strong congruence, peaking in the Middle Eocene, the Oligocene and the Plio-Pleistocene. The Eocene peak occurs when global temperature and land mammal diversity are high and is best explained as a result of niche expansion rather than abiotic forcing. Since the Eocene, there is a significant correlation between maximum size frequency and global temperature proxy. The Oligocene peak is not statistically significant and may in part be due to sampling issues. The peak in the Plio-Pleistocene occurs when global temperature and land mammal diversity are low, it is statistically the most robust one and it is best explained by global cooling. We conclude that the macroevolutionary patterns observed are a result of the interplay between eco-evolutionary processes and abiotic forcing.     

DNA sequence variation of wild barley Hordeum spontaneum (L.) across environmental gradients in Israel

Wild barley Hordeum spontaneum (L.) shows a wide geographic distribution and ecological diversity. A key question concerns the spatial scale at which genetic differentiation occurs and to what extent it is driven by natural selection. The Levant region exhibits a strong ecological gradient along the North–South axis, with numerous small canyons in an East–West direction and with small-scale environmental gradients on the opposing North- and South-facing slopes. We sequenced 34 short genomic regions in 54 accessions of wild barley collected throughout Israel and from the opposing slopes of two canyons. The nucleotide diversity of the total sample is 0.0042, which is about two-thirds of a sample from the whole species range (0.0060). Thirty accessions collected at ‘Evolution Canyon'' (EC) at Nahal Oren, close to Haifa, have a nucleotide diversity of 0.0036, and therefore harbor a large proportion of the genetic diversity. There is a high level of genetic clustering throughout Israel and within EC, which roughly differentiates the slopes. Accessions from the hot and dry South-facing slope have significantly reduced genetic diversity and are genetically more distinct from accessions from the North-facing slope, which are more similar to accessions from other regions in Northern Israel. Statistical population models indicate that wild barley within the EC consist of three separate genetic clusters with substantial gene flow. The data indicate a high level of population structure at large and small geographic scales that shows isolation-by-distance, and is also consistent with ongoing natural selection contributing to genetic differentiation at a small geographic scale.     

Non‐linearities in bird responses across urbanization gradients: A meta‐analysis

Urbanization is one of the most extreme forms of environmental alteration, posing a major threat to biodiversity. We studied the effects of urbanization on avian communities via a systematic review using hierarchical and categorical meta‐analyses. Altogether, we found 42 observations from 37 case studies for species richness and 23 observations from 20 case studies for abundance. Urbanization had an overall strong negative effect on bird species richness, whereas abundance increased marginally with urbanization. There was no evidence that city size played a role in influencing the relationship between urbanization and either species richness or abundance. Studies that examined long gradients (i.e. from urban to rural) were more likely to detect negative urbanization effects on species richness than studies that considered short gradients (i.e. urban vs. suburban or urban vs. rural areas). In contrast, we found little evidence that the effect of urbanization on abundance was influenced by gradient length. Effects of urbanization on species richness were more negative for studies including public green spaces (parks and other amenity areas) in the sampled landscapes. In contrast, studies performed solely in the urban matrix (i.e. no green spaces) revealed a strong positive effect on bird abundance. When performing subset analyses on urban–suburban, suburban–rural and suburban–natural comparisons, species richness decreased from natural to urban areas, but with a stronger decrease at the urban–suburban interface, whereas bird abundance showed a clear intermediate peak along the urban–rural gradient although abundance in natural areas was comparable to that in suburban areas. This suggests that species loss happens especially at the urban–suburban interface, and that the highest abundances occur in suburban areas compared to urban or rural areas. Thus, our study shows the importance of suburban areas, where the majority of birds occur with fairly high species richness.