1GEODIS 2.0 DOCUMENTATION 1999-2000 David Posada and Alan TempletonContact: David Posada, Department of Zoology, 574 WIDB, Provo, UT 84602-5255, USAFax: (801) 378 7423 e-mail: [email protected]. INTRODUCTIONGeoDis is program written in C and Java (two different programs that implement the same calculations)implementing the nested cladistic analysis developed by Templeton et al. (1987). Its input consists in thedescription of a nested cladogram (Templeton & Sing 1993) estimated from RFLPs or DNA sequences.The theory and applications are described elsewhere (see recommended reading). The first step is theestimation of a cladogram and the defining of a nested structure. The cladogram estimation is described inTempleton et al., (1992) and the nesting rules are described elsewhere and extended in Crandall (1996).We are currently working in the development of software for the cladogram estimation. Meanwhile, youcan find some tools to help you building the cladogram at http://bioag.byu.edu/zoology/crandall_lab/programs.htm. Outgroup probabilities (Castelloe & Templeton 1994) can also be included in the analysis.Here is a typical nested cladogram:2This cladogram consists of 21 individuals corresponding to 15 haplotype. The nested cladogram isdescribed below in the input file for GeoDis.2. INPUT FILEThe first line on the file is the name of the data set being analyzed. After that, the population informationis indicated:2.1 PopulationsThe description of the populations can be specified by their coordinates and sample size. However, in thecase of riparian or coastal species, distances are not adequately measured simply through geographicalcoordinates, and a matrix of pairwise distances among the different locations better describes thegeographical distribution in these one-dimensional habitats2.1.1 Coordinates (2-dimensions)2.1.1.1) Degrees, minutes and secondsLatitude and longitude can be specified with the standard notation degrees, minutes and seconds,followed by the letter N (North) or S (South) in the case of latitude and E (East) or W (West) in the caseof longitude. For example:23 45 00 N 34 56 78 E2.1.1.2) Decimal degreesLatitude and longitude can be also be specified as decimal degrees. In this case latitude is expressed as0-90 degrees (North {+} and South {-}), while longitude is expressed as 0-180 degrees (East {+} andWest {-}).For each population the format is:Line 1: the population number and name is specified, for example : 1 Green MountainLine 2: the sample size, the latitude and longitude are indicated, for example:7 60 22 01 N 15 20 34 E or 7 60.35 15.412.1.2 User-defined population pairwise distances (1-dimension)This information is specified as a lower triangle matrix without a diagonal (the diagonal would be madeby zeroes). The number of populations (i.e. the dimensions of the matrix, is specified above the matrix).The population number, name and size are specified at each line. The distance can be specified in anyunit. A matrix for 5 populations would look like:51 Pop-1 name Pop-1 size2 Pop-2 name Pop-2 size distance 2-13 Pop-3 name Pop-3 size distance 3-1 distance 3-24 Pop-4 name Pop-4 size distance 4-1 distance 4-2 distance 4-35 Pop-4 name Pop-5 size distance 5-1 distance 5-2 distance 5-3 distance 5-432.2 CladesThe next step in the input file is the description of the nested cladogram. Clades without geographical orgenetic variation (e.g. 1-8) are not included in the analysis. Clades at one level are subclades at the nextone (e.g., clade 1-5 is a subclade in the nested clade 2-1). 0-step clades are haplotypes. The information isspecified using the nesting clade as the unit. For each nesting clade, the composition of the clades nestedwithin is described. The clades nested within a nesting clade are denominated simply clades. Hence thespecification of cladogram starts at the 1-step level. For each nesting clade, it follows this format:Line 1 name of the nesting clade, for example Clade 1-1Line 2 number of clades nested within this nesting clade.Line 3 name of the clades nested within this nesting clade. At the nested 1-step level, the clades nestedwithin are haplotypes. We can give a name to these haplotypes, for example I, II, III,…. At higher nestedlevels (2-step, 3-step, 4-step…Total Cladogram), the name of these clades would we something like Clade1-2, Clade 2-3,…)Line 4 for each clade, its topological situation (tip = 1; interior = 0) is specified.Line 5 number of populations represented in the nesting cladeLine 6 the populations are specified by their numbersLine 7 In this line starts the observation matrix. The number of rows in this matrix corresponds to thenumber of clades specified in line 2, while the number of columns corresponds to the number of locationsspecified in line 5. For each row, and starting with the first clade (following the order specified in line 3),the number of individuals or copies of the clade is specified for each population.Line (6 + number in line 2) last line of the observation matrixThis structure is repeated for each nesting clade. After the last nesting clade (the total cladogram), in thenext line, the word "END" indicates the end of the input file.2.2.1 Outgroup weightsOutgroup probabilities for each clade can be included in the analysis (see Castelloe and Templeton 1994).If so, they have to be specified for all the clades. The outgroup weights are specified for each clade as anextra line after line 4.Line 4' For each clade, the corresponding outgroup probability is specified43. RUNNING GeoDisTo run GeoDis, the input file needs to be specified. If an output file is not specified, the results are echoedto the screen. If the C version is used, the program prompts the user for all the needed information. Forthe Java version, the appropriate checkboxes need to be specified.Number of permutationsA minimum number of 1000 permutations is recommend for a 5% level of statistical significance.4. GeoDis OUPUTThe output of GeoDis saved to a file with the same name as the input file plus the extension .out. Thevalue of the different statistics calculated is indicated for each nesting clade and its nested clades at eachlevel. Two probabilities are indicated, those corresponding to significantly small (P <=) and large values(P >= ) of the test statistic. It is highly encouraged to use the reference key in (Templeton et al., 1995) fora
View Full Document
Unlocking...