We are searching data for your request:
Upon completion, a link will appear to access the found materials.
I want to know if a given environmental variable has an effect on the range of two species/two groups of species. Specifically, I am interested in testing if one species has a "preference" for higher/lower ranges over the other (for example, do blue frogs prefer colder temperatures than red frogs?). I have occurrence data for my species as well as environmental data in a grid over the area. Can anyone point me in the right direction?
Statistical tests: which one should you use?
Published on January 28, 2020 by Rebecca Bevans. Revised on December 28, 2020.
Statistical tests are used in hypothesis testing. They can be used to:
- determine whether a predictor variable has a statistically significant relationship with an outcome variable.
- estimate the difference between two or more groups.
Statistical tests assume a null hypothesis of no relationship or no difference between groups. Then they determine whether the observed data fall outside of the range of values predicted by the null hypothesis.
If you already know what types of variables you’re dealing with, you can use the flowchart to choose the right statistical test for your data.
Understanding current and predicting future distributions of species is pivotal for ecology and for implementation of biodiversity conservation and policy measures (e.g. International Union for Conservation of Nature -IUCN Red Lists reserve selection). One of the most common methods used to gain insight in species distributions and environmental niches is Species Distribution Modelling , which is also referred to as ecological niche modelling (see discussions on terminology in , , , , ). SDM identifies locations with suitable (a)biotic conditions for species occurrences, based on climatological, environmental and/or biotic correlates . A broad range of algorithms ,  and platforms (i.e. BIOMOD, ModEco, OpenModeller, –) can be used to fit the models, each with unique features, such as different variable selecting techniques or methods for selecting (pseudo) absences –. Consequently, the best fitted model depends not only on presence data available, but also strongly on the modelling approach , . SDMs are used mainly to (1) gain insight in species’ overall distribution (i.e. , ), (2) obtain predicted occurrences for specific locations (i.e. , ) or (3) understand niche limits of species (i.e. , –). Several studies point to the need to evaluate and validate SDMs and perform in-depth analyses of the impact of algorithm selection and within algorithm consistency of predictions to generate more meaningful models , . For example, using virtual species, Saupe et al.  found that the distribution of the species data used for model training with regard to the environmental conditions available influences modelling results. Wisz et al.  showed that model accuracy (AUC values) depends on the algorithm used, reinforcing the need to assess performance of different modelling techniques , including consensus methods (that integrate the predictions of several algorithms) . Lastly, Zimmermann et al.  showed how SDM can be tailored to satisfy different aims and improve prediction accuracy. However, our screening of recent papers using SDM (see Table S1 in Supplementary material) shows that studies modelling a single species tend to use one algorithm, whereas studies modelling multiple species tend to use multiple algorithms, generally without clear explanation of the reasons for algorithms selection criteria. The 19 algorithms used in a set of 42 recent papers (Table S1) occur in both, single and multi-species studies, with Maxent (Maximum entropy) and GLM (Generalized Linear Models) being two of the most common ones. However, none of these studies analyse the advantages/disadvantages of selecting one or more algorithms, being still unclear whether species-specific features such as level of rarity, geographic spread or a combination of both, affect model fit (but see Table S1).
Here we investigate which species distribution modelling algorithms perform most consistently when: (1) evaluating overall model fit (2) evaluating spatial predictions of species occurrence at patch, landscape and regional scales and (3) identifying environmental factors as important correlates of species occurrence. We test these three aspects for a group of well-sampled hoverfly species in the Netherlands, that are selected such that they include rare to common and local to widespread species.
Important pest species of the Spodoptera complex: Biology, thermal requirements and ecological zoning
In South America, especially in Brazil, four members of the Spodoptera complex, Spodoptera albula (Walker, 1857), S. cosmioides (Walker, 1858), S. eridania (Stoll, 1782), and S. frugiperda (J.E. Smith, 1797) are important pests of many crops, in particular corn, soybean and cotton crops. Spodoptera eridania and S. frugiperda have recently invaded Africa and caused serious crop damage, and S. frugiperda has invaded Asia and Oceania. The present study tested the effect of a range of seven temperatures (18–34 °C) on these four Spodoptera species simultaneously, assessing several biological variables. Based on the thermal tolerances obtained experimentally, the ecological zoning of each species in Brazil was mapped and compared spatially, according to the crop calendar of three important crops in different regions (first and second corn harvest, soybean and cotton). Our results showed that S. eridania had the lowest temperature threshold (Tt), i.e., it is favored in regions with more moderate temperatures and did not tolerate the warmest temperature, failing to complete its development at 34 °C. In contrast, S. albula did not complete its development at 18 °C and may be more successful in warmer regions. In general, S. frugiperda and S. cosmioides were able to develop over a wide range of temperatures, and S. frugiperda showed a higher biological potential at all temperatures evaluated. Our biological data and the computational code are available online. The extensive data produced here can help other entomologists to delimit the spatial distribution of the Spodoptera complex and forecast outbreaks of these pests.
This is a preview of subscription content, access via your institution.
Multivariate approaches to determine the relationship between fish assemblage structure and environmental variables in Karatoya River, Bangladesh
Karatoya River is an important freshwater system that is home to different types of fish species. In the present study, the seasonal abundance and diversity indices of fish assemblage and their relationship with physicochemical parameters were studied. Fifty-four fish species were recorded from four stations for 12 months. The highest number of fish was recorded during the post-monsoon season and the lowest was recorded in the monsoon season. Analysis of similarity evidenced a significant difference (P < 0.05) in species abundance among the seasons. Similarity percentage analysis showed that the overall average dissimilarity among the three seasons was 73.53%, and six species were responsible for this dissimilarity. Non-metric multidimensional scaling ordination plot based on similarity indicated that the monsoon season completely differed from the other two seasons (pre-monsoon and post-monsoon) in terms of fish abundance. The values of Shannon–Wiener diversity, Margalef’s richness, and Pielou’s evenness indices varied from season to season. Canonical correspondence analysis was used to identify the most important water quality parameters influencing the seasonal variation of fish assemblage structure and to reveal the remarkable roles of alkalinity, total dissolved solids, electrical conductivity, and dissolved oxygen in the Karatoya River in Bangladesh. This study provides essential baseline information on the relationship between the environmental parameters and fish diversity in freshwater habitats in Bangladesh.
Department of Environmental Systems Sciences, ETH Zürich, 8092, Zürich, Switzerland
Alejandra Rodríguez-Verdugo & Martin Ackermann
Department of Environmental Microbiology, Eawag, 8600, Dübendorf, Switzerland
Alejandra Rodríguez-Verdugo & Martin Ackermann
Adaptation to a Changing Environment, ETH Zürich, Zürich, Switzerland
Department of Ecology and Evolutionary Biology, University of California, Irvine, CA, 92697-2525, USA
In the Keratella versus Brachionus experiment, population dynamics in the control beakers demonstrate that both Keratella and Brachionus were able to persist under both resource supply treatments (Figs 3a,b). In the competition beakers where T=0, Brachionus was driven extinct in both replicate competition beakers in 14–18 days (Figs 3c,d). Under conditions of greater variability in resource supply (T=4), Brachionus persisted in one replicate beaker, although at low and declining abundance (Fig. 3e), and went extinct in the other replicate beaker in 24 days (Fig. 3f). The dynamics of the biomass-based H′ illustrates the difference in competitive dynamics in the two treatments. H′ declined at a faster rate when T=0 than when T=4 (Fig. 4). The time taken for H′ to decline to the arbitrary value of 0.4 (using linear interpolation between time points) was used as an index of the rate of loss in community evenness. This value increased when T increased from 0 to 4 (Fig. 5a), but the trend was not statistically significant (t-test, P=0.35). In summary, increasing the temporal variability of resource supply by increasing T from 0 to 4 days did not change the outcome of competition (Keratella always dominated) and did not allow coexistence.
Population dynamics of rotifers during the Keratella versus Brachionus competition experiment. (a) & (b) show the abundance (number of rotifers per beaker) in two replicate control beakers (containing only one species) for the two food addition treatments (T=0 and 4 days). (c) & (d) show the dynamics in two replicate competition beakers for T=0. (e) & (f) show the dynamics in two replicate competition beakers for T=4. Keratella abundance=left vertical axis, Brachionus abundance=right vertical axis.
Time course of the biomass-based Shannon diversity index (H′). For the Keratella versus Brachionus experiment, individual H′ values are shown for two replicate competition beakers for each food addition treatment (T). Mean ± SE (n=3 replicate beakers) is shown for the Keratella versus Synchaeta experiment SE is not shown when only one beaker contained both species.
(a) Effect of the period of food supply (T) on the time taken for the biomass-based Shannon diversity index (H′) to decline to the arbitrary value of 0.4. For clarity, data points at T=4 have been offset horizontally. Abbreviations: B=Brachionus, K=Keratella, S=Synchaeta. (b) Effect of T on the time taken for the loosing species to be excluded (population density equals zero). Mean ± SE (where SE is larger than symbol size). For the Keratella versus Brachionus experiment, only T=0 is shown because Brachionus did not go extinct in one of the replicate competition beakers when T=4.
In the Keratella versus Synchaeta experiment, the dynamics in the control beakers demonstrate that both species were able to persist under the three resource supply treatments (Fig. 6a–c). Synchaeta went extinct in all competition beakers regardless of the level of temporal variability (Fig. 6d–l). However, temporal variability did change the dynamics of competition. Evenness (H′) declined more slowly when T=8 than when T=0 or 4 days (Fig. 4). The time taken for H′ to decline to 0.4 was the highest at T=8 (Fig. 5a) (one-way ANOVA, P < 0.01). In addition, competitive exclusion of Synchaeta took longer at T=8 than at T=0 or 4 days (Fig. 5b) (one-way ANOVA, P < 0.01). In summary, increasing the temporal variability of resource supply by increasing T from 0 to 8 days slowed the rate of competitive exclusion of Synchaeta but did not change the outcome of competition (Keratella always won) and did not allow coexistence.
Population dynamics of rotifers during the Keratella versus Synchaeta competition experiment. (a)–(c) show the abundance (number of rotifers per beaker) in control beakers (containing only one species) for the three food addition treatments (T=0, 4 and 8 days). (d)–(f) show the dynamics in three replicate competition beakers for T=0. (g)–(i) show the dynamics in three replicate competition beakers for T=4. (j)–(l) show the dynamics in three replicate competition beakers for T=8.
The experimental design of the field trials conducted in 2016 and 2017.
Plots with the different treatments of the two legume cover crop species: alsike clover (AC) and black medic (BM) was laid out in a randomized complete block design (RCBD) with Factor (1) three diversity treatments (DIV) of AC:BM (100:0, 50:50 and 0:100), Factor (2) represents three seed densities (50%, 100%, and 150% of the recommended seed density). Counting of the germinated seeds was conducted in 2016 (25 days after sowing) and in 2017 (28 DAS) on an area of 0.5 of 24 sections in 2016 (8 rows × 3 blocks) and 48 sections in 2017 (12 rows × 4 blocks).
Germination of alsike clover (AC) and black medic (BM) in monocultures and a 1:1 mixture of the two species (Mix) in laboratory experiment 2.
The seeds are germinated under four drought intensities (100%, 75%, 50%, and 25% WHC) at three temperature levels (12°C, 20°C, and 28°C). The sown seeds are 24 seed box -1 .
Partial germination ratio of alsike clover (PGRAC) and black medic (PGRBM) in a 1:1 mixture of the two species under different environmental conditions.
The different colored circles represent two field experiments (Orange circles 2016 and blue circles 2017), pot experiment (red circles), laboratory experiment 1 (white circles), and laboratory experiment 2 (green circles). The solid grey lines correspond to a GR = 1 and the broken green lines correspond to the expected PGR for the mixture. Asterisks above some data points represent a significant increase in GR >1 (P < 0.05) according to Welch’s t-test ** = P < 0.01, * = P < 0.05.
Quantitative analysis determined the occurrence of taxol in P. chienii and the variations in taxoids between P. chienii and T. yunnanensis
To date, the occurrence of taxol in P. chienii is unknown. Herein, a target LC-MS approach was applied to check the occurrence of taxol and other taxoids, such as 7-E-DAP, 10-DAB, BAC, DAP, and 7-E-PTX, in P. chienii. Our data showed that all selected taxoids were detected, confirming the occurrence of taxoids in P. chienii (Fig. 1a). Quantitative analysis revealed the variations in taxoid contents between P. chienii and T. yunnanensis. In particular, 10-DAB, PTX, and 7-E-PTX highly accumulated in P. chienii. No significant differences in the contents of BAC, DAP, and 7-E-DAP were obeserved between P. chienii and T. yunnanensis (Fig. 1b).
Determination of the contents of taxoids in P. chienii and T. yunnanensis. a A representative TIC chromatogram of six taxoids. 10-DAB: 10-deacetylbaccatin III BAC: baccatin III DAP: 10-deacetylpaclitaxel PTX: paclitaxel 7-E-DAP: 7-epi 10-desacetyl paclitaxel 7-E-PTX: 7-epipaclitaxel. b The contents of six taxoids in P. chienii and T. yunnanensis. A P value < 0.01 was considered to be statistically significant and indicated by ‘*’
Overview of the metabolites
The untargeted metabolomic analysis identified 3387 metabolites with annotations (Additional file 1). Three quality checking parameters, including total ion chromatograms, m/z widths and retention time widths, were tested, indicating that the data generated a high degree of overlap and the UPLC-MS/MS analysis reached the required standards (Additional file 2). A PCA showed that the percentages of the explained values of PC1 and PC2 were 42.12 and 18.63%, respectively, indicating a clear separation of the metabolomes from P. chienii and T. yunnanensis (Additional file 3). The metabolite profiling of the two Taxaceae species revealed great variations in their metabolomes (Fig. 2a).
The variations in the abundance of metabolites between P. chienii and T. yunnanensis. a A heatmap of the metabolites identified in the metabolomes of P. chienii and T. yunnanensis (N = 10). The heatmap scale ranges from − 2 to + 2 on a log2 scale. b KEGG analysis of all the identified metabolites. c HMDB Super Class analysis of all the identified metabolites
According to their annotations, many metabolites were assigned into different metabolic pathways. KEGG enrichment analysis showed that most metabolites were belonged to amino acid metabolism, carbohydrate metabolism, terpenoids and polyketides metabolism, and cofactors and vitamins metabolism (Fig. 2b). HMDB Super Class analysis showed that 311 metabolites were grouped to 11 major categories, such as “organic acids and derivatives” (161 metabolites), “organoheterocyclic compounds” (31 metabolites), “organooxygen compounds” (24 metabolites), “phenylpropanoids and polyketides” (23 metabolites), and “nucleosides, nucleotides, and analogue” (17 metabolites) (Fig. 2c).
Overview of the transcriptome
Using similar samples, RNA-sequencing yielded 50.35 Gb of sequence data, including 25.40 Gb from P. chienii and 24.95 Gb from T. yunnanensis (Additional file 4). The clean reads were assembled and produced 133,507 transcripts (N50: 1561), with a mean length of 513 bp, and 61,146 unigenes (N50: 1606), with a mean length of 419 bp (Additional file 5). Analysis of size distribution of all transcripts showed that 11.48% of the transcripts 11.01% of the unigenes were > 2000 bp in length (Additional file 5). For annoatation, 61,146 unigenes were annotated by several common databases (Additional file 5). The species distribution suggested that the majority of the unigenes displayed significant similarities to known proteins from Picea sitchensis, Amborella trichopoda, and Quercus suber (Additional file 5).
Analysis of DEGs showed 4,215 T. yunnanensis highly-expressed unigenes and 4,845 P. chienii highly-expressed unigenes (Fig. 3a). Most of the DEGs were assigned into different GO terms belonging to three major categories (Fig. 3b and Additional file 6). KEGG analysis showed that 34 KEGG pathways were significantly enriched in the DEGs between T. yunnanensis and P. chienii (Additional file 7). The top 20 enriched KEGG pathways, such as the ‘phenylpropanoid biosynthesis’, ‘plant-pathogen interaction’, and ‘plant hormone signal transduction’ pathways, were shown (Fig. 3c).
Identification of the DEGs between P. chienii and T. yunnanensis. a The numbers of P. chienii predominantly expressed genes and T. yunnanensis predominantly expressed genes. b GO analysis of all the DEGs between P. chienii and T. yunnanensis. c KEGG analysis of all the DEGs between P. chienii and T. yunnanensis
Variations in primary and secondary metabolism between T. yunnanensis and P. chienii
According to their annotations, a large number of DEGs were involved in the primay and secondary metabolism, and the majority of the DEGs were grouped into 46 KEGG terms belonging to 11 major categories. Significance values of each KEGG term were calculated and shown in Additional file 8. In detail, two alkaloid-related pathways, five amino acid-related pathways, two flavonoid-related pathways, one hormone-related pathway, three lipid-related pathways, one phenylpropanoid-related pathway, all three pigment/vitamin pathways, three saccharide-related pathways, one terpenoid-related pathway, and one ubiquinone-related pathway, showed significant differences between T. yunnanensis and P. chienii (Fig. 4a).
Comparative analysis of DEGs and DAMs between P. chienii and T. yunnanensis. a KEGG enrichment analysis of the DEGs. The significant P value of each KEGG term between P. chienii and T. yunnanensis was shown by a heatmap. All the KEGG terms were grouped into 11 metabolism-related categories, which were indicated by different color bars. b The relative abundances of the metabolites belonging to various major metabolic categories. c The numbers of P. chienii predominantly accumulated and T. yunnanensis predominantly accumulated metabolites in different metabolic categories
Untargeted metabolomic analysis identified 313 differentially accumulated metabolites (DAMs), 129 DAMs of which were assigned into different primary and secondary metabolite categories (Fig. 4b). The numbers of DAMs belonging to each category were shown in Fig. 4c.
Variations in the precursors of taxol biosynthesis between T. yunnanensis and P. chienii
The MEP pathway provided a key precursor, GGPP, for taxol biosynthesis . Based on the sequence similarity to model plants, a predicted MEP pathway is showed in Fig. 5a. Our transcriptome data revealed at least one unigene encoding one enzyme that is involved in the MEP pathway (Additional file 9). In the MEP pathway, three DXS encoding unigenes, one DXR encoding unigene, one MCT encoding unigene, one CMK encoding unigene, two MDS encoding unigene, two HDS encoding unigenes, one HDR encoding unigene, one GGPS encoding unigene, and three GGPPS encoding unigenes, were identified. Most of the MEP pathway-related genes highly expressed in T. yunanensis, except for CMK, GGPPS1 and GGPPS2 (Fig. 5b).
Integrated metabolomic and transcriptomic analysis of the MEP pathway. a Overview of the MEP pathway. The orange backgrounds indicated the genes identified by the transcriptome and red font indicated the metabolites identified by the metabolome. b Differential expression of the key genes involved in the MEP pathway. The heatmap scale ranges from − 4 to + 4 on a log2 scale. c Differential accumulation of the intermediate metabolites involved in the MEP pathway. “*” represents significant differences (P < 0.05)
Furthermore, our metabolome data identified four intermediate metabolites of the MEP pathway, including 2-C-methyl-D-erythritol 4-phosphate, 2-C-methyl-D-eryhritol 2,4-cyclodiphosphate, geranyl diphosphate (GPP), and GGPP. Among these intermediate products, 2-C-methyl-D-eryhritol 2,4-cyclodiphosphate, GPP, and GGPP highly accumulated in T. yunnanensis (Fig. 5c).
Variations in the taxol biosynthesis pathway between T. yunnanensis and P. chienii
Taxol biosynthesis involves a complicated metabolic pathway consisting of several intermediate products and their catalyzing enzymes . In our study, 20 unigenes encoding nine key enzymes involved in the taxol biosynthesis pathway were identified, including one TS encoding gene, four T5αH encoding genes, three TAT encoding genes, two T13αH encoding genes, five T10βH encoding genes, one TBT encoding gene, one DBTNBT encoding gene, one DBAT encoding gene and two BAPT encoding genes (Fig. 6a and Additional file 10). The BAPT1/2, DBAT, T5αH1/3, T10βH1/2/3 genes highly expressed in P. chienii and the TAT1/2/3, DBTNBT, T10βH4, T13αH1/2, TBT and TS genes greatly expressed in T. yunnanensis (Fig. 6b).
Integrated metabolomic and transcriptomic analysis of the taxol biosynthesis pathway. a Overview of the taxol biosynthesis pathway. The orange backgrounds indicated the genes identified by the transcriptome and the red fonts indicated the metabolites identified by the metabolome. b Differential expression of the key genes involved in the taxol biosynthesis pathway. The heatmap scale ranges from − 4 to + 4 on a log2 scale. c Differential accumulation of the intermediate metabolites involved in the taxol biosynthesis pathway. “*” represents significant differences (P < 0.05)
Our metabolome data identified six intermediate metabolites involved in taxol biosynthesis. Among these intermediate metabolites, 10-deacetyl-2-debenzoylbaccatin, and 10-deacetylbaccatin highly accumulated in P. chienii and 10β,14β-dihydroxytaxa-4(20),11(12)-dien-5α-yl acetate and 3′-N-debenzoyltaxol highly accumulated in T. yunnanensis (Fig. 6c).
Variations in dead-end metabolites of taxol biosynthesis and 14-hydroxylated taxoids between T. yunnanensis and P. chienii
Our transcriptome identified the encoding genes of T2αH and T7βH that are involved in the metabolism of taxusin-like metabolites and the encoding gene of T14βH that is involved in the biosynthesis of taxuyunnanin C, a classic 14-hydroxylated taxoid (Fig. 7a and b). The T2αH gene highly expressed in P. chienii, while T7βH and T14βH genes predominantly expressed in T. yunnanensis (Fig. 7c). Furthermore, several dead-end metabolites, such as (+)-taxusin, 2α-hydroxytaxusin, 7β-hydroxytaxusin and 2α, 7β-dihydroxytaxusin, and one 14-hydroxylated taxoid, taxuyunnanin C, were identified by the metabolomic analysis. The results showed that (+)-taxusin, 2α-hydroxytaxusin, and 7β-hydroxytaxusin highly accumulated in P. chienii. No significant differences in the levels of 2α, 7β-dihydroxytaxusin and taxuyunnanin C between P. chienii and T. yunnanensis were obersved (Fig. 7d).
Integrated metabolic and transcriptomic analysis of the branch of taxol biosynthesis pathway. The overview of the taxusin metabolism (a) and taxuyunnanin C metabolism (b) pathway. Orange background indicated the genes identified by the transcriptome and red font indicated the metabolites identified by the metabolome. (c) Differential expressed key genes involved in the branch of taxol biosynthesis pathway. The heatmap scale ranges from -4 to +4 on a log2 scale. (d) Differential accumulation of the metabolites involved in the branch of taxol biosynthesis pathway. “*” represents significant differences (P < 0.05)
Expression validation of the key genes involved in the taxol pathway
To investigate the differences in expression levels of the key genes involved in the taxol pathway, the relative levels of eight randomly selected taxol pathway-related genes were determined by qRT-PCR analysis. TS, DBTNBT, TAT, T13OH, T5OH genes were highly expressing in T. yunnanensis and TBT, BAPT, T10OH genes highly expressing in P. chienii (Fig. 8).
Expression validation of the key genes involved in the taxol pathway. The significant variations (P < 0.05) are indicated by ‘*’ and error bars represent mean ± SD (N = 3)
This is only the second study to thoroughly examine the intraspecific mt diversity within a single yeast species. In comparison to what was previously found, isolates sampled from L. thermotolerans harbor a set of mt genomes that are both syntenic and extremely conserved. The fact that the diversity among mtDNA in L. thermotolerans is different in comparison to the closely related L. kluyveri, but similar at the nuclear DNA level indicates that selection pressures are acting differentially on these two lineages. If they indeed occupy distinct niches, it is possible that they have varying respiration rates, which could lead to unequal dependence on the mitochondria. The ecological niche that L. thermotolerans inhabits might, in fact, also be the driving force for the difference in divergence between the mt and nuclear genomes within this species. Interestingly, it has been previously demonstrated that the environment an organism inhabits can influence mitonuclear gene complexes. Furthermore, there is evidence that changes in temperature can result in selective forces acting on gene products with disparate thermal properties (Dowling et al. 2008). Additionally, discrepancies in the reproductive cycle of these species could also have an impact on mt as well as nuclear genome evolution. Nearly all natural isolates of L. thermotolerans are haploid, whereas many L. kluyveri strains appear to be diploid or even occasionally triploid (data not shown). The lack of data concerning intraspecific diversity of mt genomes among yeast lineages is apparent. Our study revealed that evolution of the mt genomes in L. thermotoleran, and L. kluyveri is clearly different and indicates a much lower mutation rate or dramatically stronger purifying selection in L. thermotolerans. In the future, a better understanding of the forces driving these types of selective pressures will potentially reveal how closely related lineages diverge and evolve over time.