module Model where import Probability import Bio.Alphabet import Bio.Alignment import Bio.Sequence import Tree import Tree.Newick import SModel import System.Environment -- for getArgs import qualified Data.IntMap as IntMap import Probability.Logger import System.FilePath ( () ) model seqData nucs logTree = do let taxa = getTaxa seqData let n = 2 tree <- sample $ uniformRootedTree taxa (gamma 0.5 (1/fromIntegral (length taxa))) scale <- sample $ gamma 0.5 2 kappa1 <- sample $ logNormal 0 1 kappa2 <- sample $ logNormal 0 1 let tn93Model freqs = tn93' nucs kappa1 kappa2 freqs freqs <- sample $ dirichletMixture n 2 $ symmetricDirichletOn (getLetters nucs) 1 nodeMap <- sample $ iidMap (getNodesSet tree) freqs alpha <- sample $ logLaplace 6 2 let multiFreqModel = multiFrequency tree nodeMap tn93Model gammaModel = always multiFreqModel +> gammaRates alpha 4 observe seqData $ phyloCTMC tree (alignmentLength seqData) gammaModel scale let tlength = treeLength tree substs = parsimony tree (unitCostMatrix nucs) seqData addLogger $ logTree $ addInternalLabels $ scaleBranchLengths scale $ tree return ["nFreqs" %=% n, "scale" %=% scale, "scale*|T|" %=% scale * tlength, "#substs" %=% substs, "freqs" %=% sortDist freqs, "tn93:kappa1" %=% kappa1, "tn93:kappa2" %=% kappa2, "gamma:alpha" %=% alpha] main logDir = do [filename] <- getArgs let nucs = dna seqData <- mkAlignedCharacterData nucs <$> loadSequences filename logTree <- treeLogger (logDir "C1.trees") return $ model seqData nucs logTree