Sophisticated models in Bio++ Julien Dutheil, Bastien Boussau Birc, - - PowerPoint PPT Presentation

Sophisticated models in Bio++

Julien Dutheil, Bastien Boussau

Birc, Aarhus; LBBE, Lyon

Friday, December 19th 2008

• J. Dutheil, B. Boussau (Birc; LBBE)

Models in Bio++ 19/12/08 1 / 13

Models of sequence evolution

A tree

a b c

• J. Dutheil, B. Boussau (Birc; LBBE)

Models in Bio++ 19/12/08 2 / 13

Models of sequence evolution

A tree

a b c

A model of substitution

• J. Dutheil, B. Boussau (Birc; LBBE)

Models in Bio++ 19/12/08 2 / 13

Models of substitution in Bio++

• for proteins and nucleic acids (codons: soon! )
• with a gamma law to account for evolutionary rate heterogeneities

between sites

• possibility for a class of invariant sites
• possibility for covarion (heterotachous) models:
• on-off models (Tuffley and Steel 1998)
• change between rates of evolution (Galtier 2001)
• J. Dutheil, B. Boussau (Birc; LBBE)

Models in Bio++ 19/12/08 3 / 13

Homogeneous and branch-heterogeneous models in Bio++

Homogeneous model

a b c

• J. Dutheil, B. Boussau (Birc; LBBE)

Models in Bio++ 19/12/08 4 / 13

Homogeneous and branch-heterogeneous models in Bio++

Homogeneous model

a b c

Heterogeneous model

a b c

• J. Dutheil, B. Boussau (Birc; LBBE)

Models in Bio++ 19/12/08 4 / 13

A simple model of substitution: Tamura’s (1992)

• κ: Transition/transversion ratio
• θ: Equilibrium G+C content

Galtier and Gouy, Mol. Biol. Evol. 1998.

• J. Dutheil, B. Boussau (Birc; LBBE)

Models in Bio++ 19/12/08 5 / 13

Galtier and Gouy model of sequence evolution (1998)

Model

a b c

• 1 model per branch
• each model is characterized by

an equilibrium G+C content Parameters

• J. Dutheil, B. Boussau (Birc; LBBE)

Models in Bio++ 19/12/08 6 / 13

Models in Bio++

General non-homogeneous model of substitution. In the homogeneous case, θ and κ are constant over the tree (case ’a’). In Galtier and Gouy’s 1998 model, κ is constant over the tree and one distinct θ is allowed per branch (case ’b’). Between these two extrema lay models with certain branches, but not all, sharing a common value of θ (case ’c’). In the most general case ’d’, there are two sets of parameters, one for κ and another for θ, that are shared by the branches of the tree.

• J. Dutheil, B. Boussau (Birc; LBBE)

Models in Bio++ 19/12/08 7 / 13

Associating models to branches

• J. Dutheil, B. Boussau (Birc; LBBE)

Models in Bio++ 19/12/08 8 / 13

Bio++ and BppSuite

BppSuite is a set of programs implementing various methods for the evolutionary study of sequences:

• BppDist: distance estimation and tree reconstruction
• BppPars: parsimony analyses
• BppML: ML reconstruction of phylogenetic trees, including

using non-homogeneous models

• BppSeqGen: sequence simulation, including using

non-homogeneous models

• BppAncestor: ancestral sequence reconstruction, including

using non-homogeneous models

• BppSeqMan: sequence and alignment manipulation
• BppConsense: building of consensus trees
• BppPhySamp: select sequences according to a tree or a distance

matrix

• BppReRoot: automatic re-rooting of trees
• J. Dutheil, B. Boussau (Birc; LBBE)

Models in Bio++ 19/12/08 9 / 13

Specifying options of BppSuite programs

Launching an analysis with bppml Example: bppml param=fichier.opt fichier.opt alphabet = DNA sequence.file = sequences.fasta sequence.format = Fasta sequence.sites to use = complete tree.file = tree.dnd etc...

• J. Dutheil, B. Boussau (Birc; LBBE)

Models in Bio++ 19/12/08 10 / 13

Associating models to branches in BppSuite

• J. Dutheil, B. Boussau (Birc; LBBE)

Models in Bio++ 19/12/08 11 / 13

Exercise

• THE DATA: A well-known scientist is working on a family of

homologous genes (file ”sequences.fasta”). These sequences come from closely-related species and have been named according to their species of origin: S vulg, S con, S dio, S lat, S dic. Specifically in species S dio, S lat, and S dic, the gene is found on sexual

• chromosomes. For each of these species, the alignment thus contains

two sequences, one from the X chromosome (X is put at the end of the name), and one from the Y chromosome (Y is put at the end of the name). The famous scientist has built a rooted phylogenetic tree relating all sequences in his dataset (file ”tree.dnd”).

• THE PROBLEM: The scientist suspects there might have been some

Biased Gene Conversion (BGC) going on on the branch leading to the group containing sequences S dioY, S latY, and S dicY. This BGC is expected to increase the number of substitutions towards bases G and

• C. Your aim is to test for the presence of BGC on this branch.
• J. Dutheil, B. Boussau (Birc; LBBE)

Models in Bio++ 19/12/08 12 / 13

Exercise

• THE AIMS:
• Using bppML, devise a test to see whether the data rejects BGC on

this branch.

• Meanwhile, try to accurately characterize the evolution in this dataset.

Is there significant rate heterogeneity? Covarion-like evolution? How important has been process heterogeneity in the evolution of this dataset?

• THE METHOD:
• Option files have been partially filled. You need to complete them to

build a proper model to make hypothesis 0 (model 0: there was no heterogeneity in the evolution of the dataset), hypothesis 1 (model 1: there has been one significant change in the evolutionary process on

• ne particular branch), hypothesis 2 (model 2: the evolution has been

globally heterogeneous, with different processes on different branches).

• Play with the options to better characterize sequence evolution
• Use likelihood ratio tests to compare hypotheses.
• BONUS QUESTION:
• Think of another way to test whether the evolutionary process has

been particular on the branch of interest. BppSuite may be useful once again; you may need to do a little bit of programming.

• J. Dutheil, B. Boussau (Birc; LBBE)

Models in Bio++ 19/12/08 13 / 13