The course provides an introduction to likelihood-based inference in Biology. We will cover both theoretical and practical aspects of maximum likelihood and Bayesian inference.
Meeting: Monday and Wednesday 10:00-10:50AM in 2025 Haworth
Instructors (office hours by appointment):
|John Kellyemail@example.com||864-3706||5005 Haworth|
|Mark Holderfirstname.lastname@example.org||864-5789||6031 Haworth|
Grades will be based on class participation and homework assignments. We will have approximately one homework assignment per week.
A free-of-charge installation of python with the packages that we will use are available from
Anaconda from Continuum Analytics.
RStudio is useful if you prefer to use R.
There are some difference between a DOS prompt and UNIX. For the simple stuff, the main difference is
dir on Windows instead of
ls on Unix.
You probably should check out the "Intro-to-Unix" section of the Woods Hole Workshop on Molecular Evolution's Computer lab introduction
if you have never worked from a shell before.
|Date||Topic and Links||Assignments|
|Week 1 Jan 20||Probability, random variables, distributions
Notes on probability by Bálint Tóth. Some of MTH's notes on JKK's lecture
|Week 2 Jan 25, Jan 27||Random samples, sample distributions, likelihood. Some of MTH's lec2.pdf and lec3.pdf on JKK's lectures||HW 1 (due Wed., Feb 3). A (corrected) spreadsheet with the data for the birds is available as a zip archive: Jack.ods (or Jack.xls). You will also need PoissonCounts.xls
Laura's as a HW1_Laura.pdf
|Week 3 Feb 01, Feb 03||Explicitly specifying variability: likelihood examples. MTH's lec4.pdf and lec5.pdf notes. JKK's saxony.ppt||SimpleLobsters.xls
HW2_2016.doc revised (due Wed. Feb. 17)
|Week 4 Feb 08, Feb 10||Maximum likelihood estimation and Markov chains. MTH's lecture 6 notes and lecture 7 notes|
|Week 5 Feb 15, Feb 17.||Likelihood ratio test statistic and Model Selection lec8.pdf twoParamNormal.pdf||HW3_2016.doc
|Week 6 Feb 22, Feb 24||Generalized Linear Models (continued)|
|Week 7 Feb 29, Mar 04||Likelihood ratio test statistic and Model Selection||numerical optimization notes
1 param optimizer: one-param-k2p-py.txt
2 param optimizer: k2p-py.txt
|Week 8 Mar 07, Mar 09||Logistic regression||JKK's code for logistic regression
data for JKK's code
MTH version with parametric bootstrapping.
|Week 9 Mar 21, Mar 23||Bayesian methods||hw4.pdf
Notes on MCMC for Bayesian inference and the coin_contamination.py script
|Week 10 Mar 28, Mar 30||Computational stuff||high-level.py - the start point for class on Monday
work-in-progress.py script that I worked on in class.
|Week 11 Apr 04, Apr 06||Computational aspects: Markov chain Monte Carlo (MCMC) hastingsRatio.pdf||hw5.pdf, hw5template.py and hw5-data.fas.txt
continuous-mcmc.py and diagnostics.R and summarize.R. See Patrick Lam's very nice slides
|Week 12 Apr 11, Apr 13||Multiparameter MCMC||MCMCMC version of the continuous MCMC example|
|Week 13 Apr 18, Apr 20||Hastings ratio and model jumping||hw6.pdf (with links in that doc) and Bayes Factor slides Green's hastings ratio recipe. Wed: model-jumping-mcmc.py. and hw5-answer.py|
|Week 14 Apr 25, Apr 27||HMM||JKK's hmm1.py and hmm2.py|
|Week 15 May 02, May 04||Finishing model jumping. Ascertainment bias (maybe)||likelihood-hw6-mth-answer.py (not tuned at all, and only doing a sliding window proposal). And likelihood-hw6-mth-answer-opt-1.py (an optimized version that runs about 350 times faster on my machine.