evotext/ielex-data-and-tree: IELex data and tree (2021/11/08)

Data and scripts for producing a Baysian phylogenetic tree sample of for the Indo-European family which is "good enough", using data from IELex (Dunn et al. 2011).

This repository contains data and scripts for producing a Baysian phylogenetic tree sample of for the Indo-European family which is good enough for use in phylogenetic comparative methods. It should be considered as the current version of the "Indo-European Lexical Cognacy Database" from Dunn et al. (2011), popularly known as "IELex". The "Indo-European Cognate Relationships" (IE-CoR) database project, now based at the Max Planck Institute for Evolutionary Anthropology, has been working on a much improved database originally based on this data, but has not had any public data release so far. In the meantime, this is the best we can offer, as an alternative to the original website (http://ielex.mpi.nl) which is now offline.

The data/ directory holds the files most researchers will be interested in:

ielex.csv is a single table in long format containing the essential database information, given one entry per row, each with the associated language, concept, and cognate set. It also carries the glottocode (Hammarström et al., 2021) corresponding to each language; concept glosses are given via the corresponding Concepticon (List et al., 2021) cognate set. It is the file most people will want when searching for "IELex data".

concepts.csv is a list mapping all concepts used in IELex to their corresponding concept set in Concepticon (List et al., 2021), giving both the gloss and the id.

ielex.nex is a NEXUS file with the information from the table above, with ascertainment correction columns, charstate labels, and assumptions. The file is built with build/build_nexus.py.

ielex.nn.pdf and ielex.nn.png are a PDF and a high-resolution PNG from the NEXUS file above, generated with SplitsTree (Huson & Bryant, 2006).

ielex.mcc.tre is the Maximum Clade Credibility ("consensus") tree from our latest analysis (currently IE-tree-v1). The first 50% of the sample was removed as burn-in.

ielex.mcc.pdf is a graphic rendering of the tree above, generated with FigTree.

The build/ directory holds the files used for preparing the phylogenetic reconstruction.

The IE-trees-v1/ directory holds the files related to the phylogenetic reconstruction, including model, logs, and state. As stated above, a summary of the main output, as ielex.mcc.tre, is found in the data/ directory. For statistical purposes researchers are likely to want the entire (unsummarised) tree sample, IE-trees-v1/ie-v1.nex.