Phylogenetics: Syllabus
EEB 5349: Phylogenetics | |
Lectures: TuTh 11:00-12:15 (TLS 313) Lab: Th 2-4 (TLS 313) Lecture Instructor: Paul O. Lewis Lab Instructor: Daniel Fan |
Lecture Topics
The following syllabus is tentative and probably will change without notice numerous times during the semester. Also, the content of linked presentations may change as well (so if you intend to print out lectures before class, do so as late as possible). Changes made after lectures are given will primarily reflect correction of typographical errors. Downloading PDFs from this web site now requires a username and password. All content linked to this page is copyright © 2011 by Paul O. Lewis.
Day | Lecture | Lab/Homework |
Tue., Jan. 18 | Introduction The terminology of phylogenetics, rooted vs. unrooted trees, ultrametric vs. unconstrained, paralogy vs. orthology, lineage sorting, "basal" lineages, crown vs. stem groups |
Homework 1: Trees from splits |
Thu., Jan. 20 | Introduction to optimality criteria and search strategies Exhaustive enumeration, branch-and-bound search, algorithmic methods (star decomposition, stepwise addition, NJ), heuristic search stragegies (NNI, SPR, TBR), evolutionary algorithms |
(1) Nexus data file format, (2) using the cluster, and (3) Introduction to PAUP* |
Tue., Jan. 25 | Consensus trees , Parsimony and History of Parsimony (PL away: watch Consensus.mov, Parsimony.mov and ParsimonyHistory.mov) Strict, semi-strict, and majority-rule consensus trees; maximum agreement subtrees; Camin-Sokal, Wagner, Fitch, Dollo, and transversion parsimony; step matrices and generalized parsimony; History of parsimony: Hennig, Edwards, Sokal, Camin, Dayhoff, Kluge, Farris, Fitch, Sankoff, and Wiley; character vs. character state. |
Homework 2: Parsimony |
Thu., Jan. 27 | Bootstrapping and Distance Methods (PL away: watch Bootstrapping.mov and Distances.mov) Bootstrapping; Distance methods: split decomposition, quartet puzzling, neighbor-joining, least squares criterion, minimum evolution criterion |
Snow day, no lab |
Tue., Feb. 1 | Substitution models (watch ModelsIntro.mov if PL not back yet) Transition probability, instantaneous rates, JC69 model, K2P model, F81 model, F84 model, HKY85 model, GTR model |
Homework 3: Distances |
Thu., Feb. 3 | Substitution models Continue discussing models, and/or use the time to answer any questions that have built up while I was away. |
Searching (start Python 101 if there is time) |
Tue., Feb. 8 | Maximum likelihood Poisson processes; Likelihood: the probability of data given a model, maximum likelihood estimates (MLEs) of model parameters, likelihood of a tree, likelihood ratio test |
Homework 4: Likelihood |
Thu., Feb. 10 | Rate heterogeneity Proportion of invariable sites, discrete gamma, site-specific rates |
Likelihood |
Tue., Feb. 15 | Amino Acid, codon and secondary structure models Empirical amino acid rate matrices, transition probabilities by exponentiating the rate matrix, RNA stem/loop structure, compensatory substitutions, stem models, nonsynonymous vs. synonymous rates, codon models |
Homework 5: Rate heterogeneity |
Thu., Feb. 17 | Model selection Likelihood ratio test (LRT), Akaike Information criterion (AIC), Bayesian Information Criterion (BIC) Expected number of substitutions for a model An example calculation for the F81 model |
Python 101 |
Tue., Feb. 22 | Long-branch attraction Statistical consistency, long branch attraction (real), long branch repulsion (real?) Simulation How to simulate nucleotide sequence data, and why it's done |
Homework 6: Simulation |
Thu., Feb. 24 | Statistical tests involving phylogenies ILD parsimony test for combinability, KH test, SH test, SOWH test |
ML analyses of large datasets |
Tue., Mar. 1 | Bayes primer' Conditional/joint probabilities, Bayes rule, prior vs. posterior distributions, probability mass vs. probability density, Markov chain Monte Carlo |
No homework this week |
Thu., Mar. 3 | Bayes primer (continued) | Midterm exam |
Tue., Mar. 8 | Spring break | no class |
Thu., Mar. 10 | Spring break | no class |
Tue., Mar. 15 | Discussion of midterm Bayes primer (continued) Phylogenetic applications of Markov chain Monte Carlo Prior distributions Commonly-used prior distributions |
no homework this week |
Thu., Mar. 17 | Prior distributions Commonly-used prior distributions (continued) Confidence vs. credible intervals Frequentist confidence intervals differ from Bayesian credible intervals |
Using R to explore probability distributions |
Tue., Mar. 22 | Priors (cont.) (error corrected on slide 27) Pros and cons, hierarchical models and hyperpriors |
Homework 7: MCMC |
Thu., Mar. 24 | Bayesian model selection (updated to include stepping-stone slides) Bayes factors, posterior predictive approaches to model selection |
MrBayes lab |
Tue., Mar. 29 | Star tree paradox (updated to include a few slides on BIC) When posteriors and bootstraps conflict |
Homework 8: Larget-Simon move |
Thu., Mar. 31 | Models for discrete morphological data DNA sequences vs. morphological characters, Symmetric vs. asymmetric 2-state models, Mk model, Tuffley-Steel no-common-mechanism model |
Morphology and partitioning in MrBayes |
Tue., Apr. 5 | The No-Common-Mechanism model (updated 4/5/11) The likelihood model that behaves like parsimony Discrete character correlation Pagel's likelihood ratio test Continuous character correlation Felsenstein's independent contrasts |
Homework 9: Independent contrasts (assigned Thursday) (corrected) |
Thu., Apr. 7 | Ancestral states Likelihood, (empirical) Bayesian and parsimony reconstruction of ancestral states |
Phycas |
Tue., Apr. 12 | Stochastic Character Mapping Concentrated changes test, stochastic mapping for estimating ancestral states and character correlation, SIMMAP demo |
Homework 10: Stochastic Mapping |
Thu., Apr. 14 | Mixture models rjMCMC, heterotachy models, Dirichlet process prior models |
BayesTraits lab |
Tue., Apr. 19 | Divergence time estimation (part 1) Non-parametric rate smoothing, penalized likelihood, cross-validation |
no homework, work on project |
Thu., Apr. 21 | Divergence time estimation (part 2) Bayesian approaches: Thorne/Kishino autocorrelated log-normal model; BEAST uncorrelated log-normal model; coalescent, exponential growth coalescent, and Yule tree priors |
BEAST lab |
Tue., Apr. 26 | Lineages through time analyses | no homework, work on project |
Thu., Apr. 28 | Estimating species trees Evaluations (last part of lecture) |
APE Lab |
Wed., May 5 | Final exam and project reports due by 5:30pm (but if you need an extension, write to me) |
Goals of this course
This course is designed to give you the background you need to understand and critically evaluate phylogenetic analyses described in current primary literature, and to design appropriate phylogenetic analyses to address your own research questions.
Compared to many graduate courses, you will spend less time reading papers and more time using state-of-the-art software tools and doing homework assignments designed to ensure that you understand the output of the programs.
There is a confusing diversity of programs these days for performing phylogenetic analyses. We will concentrate on only a few so that you will know how to use these well by the end of the course.
Textbook
No textbook is required for this course, although you might find Joe Felsenstein's 2004 book "Inferring Phylogenies" (published by Sinauer) useful.
Labs
The laboratory section of this course consist of tutorials that you work through at your own pace using your own laptop computer. In some cases, you will use the UConn Bioinformatics Facility's computing cluster to perform analyses. Please contact Jeff Lary (486-5036) to get an account on the cluster at your earliest convenience.
Homeworks
Your grade will be based on a midterm exam, a final exam and a number of homework assignments, one of which will be assigned (nearly) every week. These homework assignments should be treated as if they were take-home, open-book exams. You may therefore consult with either me or the TA for the course, but not with fellow students when working on the homeworks.
Projects
In addition to homeworks, you will prepare a term paper to be due the last week of the course. There is a lot of flexibility in the nature of the term paper. If you have data of your own, you may decide to write a paper describing a phylogenetic analysis of these data, using appropriate methods learned during the course. If you are not yet at the stage of your graduate career where you have data of your own, you can do a thorough re-analysis of an existing data set. Finally, it is ok to simply write a review paper describing a particular topic in phylogenetics in depth. Please get my approval of your chosen topic before doing extensive work on your paper.