APE incomplete distance extension
Abstract: Constructing phylogenetic trees from measures of dissimilarity (distances) between species is a popular approach for inferring phylogenies, as it is fast and provides a useful snapshot of the data. The most popular method for such reconstruction is Neighbour Joining (N.J). Many improvements have been made to this algorithm since its release, for instance Bio-NJ, which have sped NJ up and have improved its accuracy. However in practice distance matrices with missing values (due to noise or lack of confidence in readings) are encountered. For such cases modifications of the above methods have been made, such as NJ*, which given an incomplete distance matrix will return a phylogenetic tree from that distance matrix. The project aims to extends the APE R package (which currently contains implementations of NJ and Bio_NJ) to also include such methods for phylogenetic reconstruction. We will also add a non-agglomerative approach (the triangles method) to APE and later extend this to handle reconstruction from incomplete distances.
Here is a week-by-week project schedule.
May 24th to May 31st