The selection of sequences had to be undertaken manually due to a high error rate in the names and information fields provided for the publicly deposited entries. In addition, among multiple entries for a single type strain, the best quality sequence was selected for the project. The living tree database provides corrected entries and the best quality sequences with a manually checked alignment. In total, with LTP release 100, 750 new distinct species (table1.pdf) have been incorporated into the database and tree. The tree reconstruction has been performed by using the maximum likelihood algorithm RAxML. The tree provided in the current release is a result of the calculation of a dataset containing 10,950 single entries, 7,710 corresponding to type strain gene sequences, as well as 3,240 additional high quality sequences to give robustness to the reconstruction.

Trees are dynamic structures that change on the basis of the quality and availability of the data used for their calculation. Therefore, the addition of new type strain sequences in subsequent releases may help to resolve certain branching orders that appear ambiguous in this first release.