MODULARIZACE VÝUKY EVOLUČNÍ A EKOLOGICKÉ BIOLOGIE CZ.1.07/2.2.00/15.0204 PF_72_100_grey_tr ubz_cz_black_transparent http://ars.els-cdn.com/content/image/1-s2.0-S1360138510000981-gr2.jpg http://www.flywings.org.uk/MorphoMeet12/images/skull_lmks.jpg http://2.bp.blogspot.com/-QQBb7pH-BFU/UBzc53AZIaI/AAAAAAAAFL4/-SA2VMszaL4/s1600/zuttiyeh.png http://host.uniroma3.it/laboratori/paleontologia/rhino2.jpg http://www.pnas.org/content/104/42/16604/F2.large.jpg http://www.flywings.org.uk/images/smallCPKSignet.JPG Genetic methods and morphology What is the genetic basis of a morphological trait? (quantitative trait loci = QTL) Trait variation in time (phylogenesis) Trait variation depending on other factors Some methods are shared (e.g. PCA) Pečnerová et al. Syst. Biol. (2015) DNA morfometrie Molecular vs. morphological traits amount (103 - >106 vs. 102) independence phylogenetic scale (with mol. traits we can compare e.g. bacteria and vertebrates) larger number of taxa usually represent many genes (vs. e.g. mtDNA, cpDNA) we can also study museum/fossil material traits variation genetic determination qualitative discrete one to a few genes of large effects quantitative – plastic continuous many genes of small effects + non-genetic influence quantitative – meristic continuous scale of discrete traits many genes of small effects + non-genetic influence (threshold traits) Many so-called qualitative traits have, in fact, quantitative basis! •qualitative traits • •epigenetic traits • •traditional morphometrics • •geometric morphometrics Analysis of phenotype http://www.softchalk.com/lessonchallenge09/lesson/genetics/Mendel_and_pea_plants_final.jpg Mendelian inheritance, 1 - a few genes mutations in D. melanogaster mutations of Hox genes: Antennapedia, Ultrabithorax Qualitative trais http://www.scielo.br/img/revistas/isz/v98n3/a09fig02.jpg colouration: scarlet tiger moth (přástevník hluchavkový, Callimorpha dominula), grove snail (páskovka hajní, Cepaea nemoralis), beetle elytrons mammals: ~15 domestic and laboratory species - cat, mouse, guinea pig, weasel, leopard, mink, horse pigments: eumelanin, phaeomelanin, carotens, haemoglobin, trichosiderin Callimorpha dominula - přástevník hluchavkový Cepaea nemoralis - páskovka hajní http://www.vivarium.cz/Image/Savci/Zvirata/morcata.jpg http://4.bp.blogspot.com/-hGrfB8jGHYY/TgLC9U1OxbI/AAAAAAAAAkk/uIeGa5mLMXY/s400/IMG_6370.JPG http://www.chovatelka.cz/system/files/photos/2001/images/original.jpg?1322075554 Qualitative trais main allelic series – mouse: A = agouti (colour structure along hair) B = brown (protein component of pigment granules) C = albino (reduction of number of pigmented lesion) D = dilute (aggregation of pigmented lesions) E = extension (changes in amount of eumelanin) DBA = dilute–brown–non-agouti musculus http://images.sciencedaily.com/2008/08/080815140250-large.jpg merle1.jpg dilute albino agouti Qualitative trais main allelic series – cat: A-agouti, T-tabby, B-brown, O-orange, S-white spotting, W-white, L-long hairs http://4.bp.blogspot.com/_cOYt1cVgkNw/S8TUZeKsoSI/AAAAAAAAABY/_thmaLA31YQ/s1600/normal_whiskas_cat_ 11.jpg http://loyalsnowclan.webs.com/Cindercoal.jpg http://i3.squidoocdn.com/resize/squidoo_images/250/draft_lens2340444module13089432photo_1229573686b lackcat.jpg http://i2.squidoocdn.com/resize/squidoo_images/250/draft_lens2340444module13275619photo_1230904300o rangecat.jpg http://i2.squidoocdn.com/resize/squidoo_images/250/draft_lens2340444module13089240photo_1229572704w hitecat.jpg Qualitative trais Epigenetic traits •Epigenesis = developmental interactions over/outside of alteration of genes • •basic criterion = absence of correlation between the trait and its size Epi 1_1 •Relationship of genotype and phenotype: •VP = VG + VE –VP = total phenotype variance –VG = genotype variance –VE = variance caused by environment –VG = VA + VD + VI A = additivity; D = dominance; I = epistasis – Quantitative traits Heritability, h2: = measure of heritable part of phenotypic variability says, to what extent phenotypic variance has genetic basis True heritability: in narrow sense h2 = VA / VP in broad sense h2 = VG / VP Quantitative traits IW ZW BCW OL RW IOCW Traditional morphometrics Principal components analysis (PCA) reduction of data dimensionality with as low information loss as possible exploratory data analysis making predictive models n individuals p variables correlation or covariance matrix eigenvalue = latent root (latentní kořen) eigenvector = latent vector (latentní vektor) PC1: vysvětluje největší podíl variability PC2: druhý největší podíl variability atd. loadings (zátěže) PC scores PC1 PC2 Kaiser’s criterion: 5 PCs scree graph: 3–4 PCs size vector how many components? •Problem of multiple groups: •CPCA (common PCA) •MGPCA (multiple-group PCA) PC1 PC2 CPCA MGPCA same variation same direction different variation same direction different variation different direction 1xx Problem of size: omitting PC1 Burnaby’s adjustment Slované Blízký východ Britské ostrovy Středomoří Kavkaz Z Asie Írán Baskové aDNA současní lidé aDNA kultura zvoncovitých pohárů •Multidimensional scaling (MDS): •not only correlation/covariation matrix, all types of matrix, e.g. similarity/dissimilarity matrix • •Classical MDS = principal coordinates analysis (PCoA) •Metric MDS •Non-metric MDS •Generalized MDS •Discriminant function analysis (DFA) and canonical analysis (CVA): •a priori groups •minimization of within-group variation •maximization of among-group variation •Mahalanobis (generalized) distances •MANOVA (Wilk’s Lambda, Pilai’s trace) •Hotelling T2 test •stepwise DFA • •Cluster analysis 1_11 Minimum spanning tree (MST) 1 3 2 6 4 8 7 5 CV1 CV2 A. Dürer (1524): Vier Bücher von Menlicher Proportion. Durer Geometric morphometrics Darcy W. A. Thompson (1917): On Growth and Form Absence of quantification of shape changes! In the past, there were two different strategies in study of shape of biological objects: 1. W. D’Arcy Thompson 2. Traditional morfometrics: F. Galton, K. Pearson, R.A. Fisher, S. Wright, H. Hotelling... linear measurements, weights, angles, surfaces... PCA, DFA, CVA, FA, PCoA, cluster a. Absence of any information on shape (morphometrics)! In the past, there were two different strategies in study of shape of biological objects: Geometric morphometrics I. Analysis of closed curves Zub2 Fourier1 cos sin harmonics, coefficients Fourier analysis Fourier2 Traditional Fourier a. Zub2 x y Elliptic Fourier a. F1 F2 F3 F4 F5 Error landmarks quantification of shape using shape coordinates, distinction between different shape components information on shape maintained during the whole analysis size standardization and ability to work independently with the size vector data processing with traditional morphometric methodology Geometric morphometrics II. Analysis of landmarks landmarks = points that can be accurately localized and which are – at least in a geometric sense – homologous among the objects Location of 15 homologous landmarks used for geometric morphometrics to... | Download Scientific Diagram The 45 landmarks used in the geometric morphometric analysis, and the... | Download Scientific Diagram shape = everything except information on size, position, and orientation of objects Procrustes superposition = GLS (Generalized Least Squares) The eiffel tower in paris, france plakáty na zeď • plakáty dovolená, cestování, věž | myloview.cz The eiffel tower in paris, france plakáty na zeď • plakáty dovolená, cestování, věž | myloview.cz The eiffel tower in paris, france plakáty na zeď • plakáty dovolená, cestování, věž | myloview.cz 1) Size adjustment: unit centroid size 1) Size adjustment: unit centroid size 2) Transformation: same centroid position 3) Rotation: minimization of distances between homologous landmarks SHAPE SPACE: n = pk – k – k(k–1)/2 – 1 shape coordinates 3) Rotation: minimization of distances between homologous landmarks Extracting shape information: Procrustes superposition 1. Change scale so that all configurations have the same size 2. Superposition of the centers of gravity on a single point 3. Rotation to minimize the dispersion of corresponding points Original landmark configurations T C A A´ tangent space shape space tangent (reference, consensual) configuration 1_18 metaphor of infinitely large and infinitely thin metal plate Thin-Plate Spline (TPS) http://www.virtual-anthropology.com/virtual-anthropology/images/GMM_grid.jpg energy necessary for plate deformation = bending energy differentiating between afine and nonafine shape changes projection of latent roots of bending energy into components = partial warps partial warp 0 ~ afine component Thin-Plate Spline (TPS) Afine shape change Afine shape change Afine shape change parallel lines are still parallel Afine shape change Nonafine shape change parallel lines are not parallel any more Nonafine shape change Homo reference object Pan reference object = Homo reference object = Pan RW 1 RW 2 >2 samples: Thin-Plate Spline Relative Warps (TPSRW) ~ PCA RW 1 RW 2 RW 1 RW 2 http://life.bio.sunysb.edu/morph/ Software: tpsDig: úprava obrázků, digitalizace bodů, měření rozměrů tpsSplin: TPS tpsRelw: TPS Relative Warps tpsRegr: regrese na nezávislou proměnnou tpsPLS: metoda parciálních nejmenších čtverců (např. korelace 2 sad bodů) tpsSuper: deformace obrázků („unwarping“) tpsTree: analýza tvarových změn podél větví fylogenetického MorphoJ http://www.flywings.org.uk/MorphoJ_page_files/MorphoJ_logo.jpg 9.17.jpg 9.18.JPG Morfometrics and phylogenesis tpsTree mapping of shape changes Landmark based methods without landmarks – „sliding semilandmarks“ Bookstein et al., Anat. Record (1999) http://38.media.tumblr.com/407876e87b323fb462fc358fb659f7a0/tumblr_n3qyuoyjB41r46foao1_1280.png http://40.media.tumblr.com/afc72807c1db7dc44ae5b05414789b20/tumblr_n3qyuoyjB41r46foao3_500.png http://a.fsdn.com/con/app/proj/morpho-rpackage/screenshots/pcaplot.png Velemínská, Dupej, Živa 5/2016 Velemínská, Dupej, Živa 5/2016 Velemínská, Dupej, Živa 5/2016 Example: knockout of gene Sprouty Program Checkpoint (Stratovan): Problem of symmetric objects Problem of symmetric objects Problem of symmetric objects matching symmetry Klingenberg et al., Evolution (2002) matching symmetry Klingenberg et al., Evolution (2002) object symmetry object symmetry Comparison of wild-type and mutant mouse, 5 weeks red = mutant blue = wild-type Stratovan Checkpoint MorphoJ 9.19.JPG 9.20.JPG 9.21.JPG Comparative analysis 9.23.JPG Independent contrasts assumption: Brownian motion! Alternatives: phylogenetic generalized least squares (PGLS) possibility of also applying other models than Brownian motion 9.9.JPG 9.10.JPG Brownian motion model Ornstein-Uhlenbeck model 9.12.JPG