A repository of datasets that are set up to be used with PartitionFinder (www.robertlanfear.com/partitionfinder), which is a program for performing model selection and partitioning scheme selection.

All of the datasets are publicly available. Each folder here contains a published alignment of DNA or amino acid sequences, with information about the original study.

It's a place to store and collect datasets for which model selection might be an issue. In phylogenetics, we often go to some lengths to choose a model of evolution that minimises some information criterion like the AICc. This repository should make it easier to develop new methods and compare them to old ones, because it contains a well curated collection of datasets which have been, or could be, used to test model selection algorithms.

Inside each folder is:

1. A README file which contains the reference to the original paper and/or the dataset itself, plus any additional comments on the data (e.g. Taxon, number of spp, number of sites, etc).

2. An alignment file in phylip format. This will be the original alignment described in the README, converted if necessary into phylip format.

3. A partitionfinder.cfg file which describes the dataset for a PartitionFinder. Most of the settings in this file are arbitrary, and are all identical. The only settings that are not arbitrary/identical are the alignment name, and the description of the data blocks.

All of the datasets in this file are publicly available, and links to the original data, or descriptions of how it was obtained, are provided in the README file of each folder.

Yes. All of the original datasets are publicly available and can be re-used. If you use any of the datasets, you MUST cite each original study, and the original datasets if they have their own DOIs. I provide copies of them here, as well as links to the original source of the data, and references for the original publication and the dataset if it has a separate reference.

I DO NOT own the copyright to these datasets, but all of them are in the public domain and some have their own licences.

Everything here that is not a dataset (the partitionfinder.cfg files and the README in each folder) is released under a CC-BY 3.0 license: http://creativecommons.org/licenses/by/3.0/deed.en_US) full text here: http://creativecommons.org/licenses/by/3.0/legalcode

If you use any of the datasets here, make sure to reference three things:

1. The original study,

2. The dataset itself if the dataset has its own DOI (e.g. data dryad datasets have their own DOI's and should be referenced separately).

3. This repository (github.com/roblanf/PartitionFinder_datasets)

This is essential to reward and acknowledge those who spend weeks and months in the field, laboriously chasing frogs/flied/lizards etc., then are kind enough to share their data with the world so that people like me (and you, if you're reading this) can re-use them for other things.

If you have any questions, or would like your dataset to be included here, or removed, please contact me. Contact details are at www.robertlanfear.com/contact

Rob Lanfear March 2013