Updating manuscript

The progress of the past few days has lead to substantial revision proposals in my manuscript draft. Rewriting for discussion with Brian tomorrow. Currently revised outline:

  1. Intro
    1. Biological Motivation
    2. Comparative Methods and Comparing Models
    3. Model choice paradigm: the right direction, but incomplete
  2. Simple Example
    1. The Star Tree
    2. Two hypotheses: stabilizing vs disruptive selection
    3. Information Criteria approach is insufficient
    4. Likelihood surfaces
    5. Likelihood ratio distribution
  3. Anoles Example
    1. Background to example and models
    2. Initial conclusions from Butler & King
    3. Raising skepticism – simulated data sets
    4. A robust approach: parametric bootstrap
  4. Power of a Tree
    1. Quantifying information content of a tree a priori
    2. independent of data
  5. Discussion
    1. Model choice paradigm
    2. Need for power analysis
    3. Future of model comparisons in comparative methods

Original draft is in phylotrees/writeups/note_on_aic.tex preserved in revision 205 of the svn repository (in case I’m trying to refer to some of that draft text later). Also have copies with comments from Chris, Matt, and Roi.

Repository Updates

Reorganized the svn repository for the project. Trying to keep a standard structure and version manage all contents. Current structure is:

  • project
    • articles
    • writeups
    • code
    • images
    • data
    • presentations

articles are imported into My Mendeley Library, which keeps its own master copies so the articles section here is no longer needed. Writeups are drafts of papers, relevant images are in the images directory which may need an image database manager such as picasa as it gets large. Images I’m unlikely to need can be deleted and the whole directory can be kept under version management. Similarly the data folder can also be version managed, even though not getting regular updates. Having the entire project under svn makes the working directory for a project expendable, and easy to keep only active projects. Can always look into the svn repository to checkout a copy of an old project or part of one.