LOFT TREE FORMAT 05-01-2006 (DDMMYYYY) With respect to the syntax of the tree files, LOFT follows the NHX format, extends somewhat to that and is a bit more flexible w.r.t. current conventions. As NHX, it requires at least TWO BRANCHES of each internal node. However, it does allow empty nodes in the tree file, while assigning unique nodenames when none is specified. Comments and NHG tags are only allowed after (internal) nodenames (and optionally their accompanying optional branchlengths). commenst are not allowed to span multiple lines, except between NHX tags. Quoted strings are allowed but are taken literally, i.e. including their quotes. Quoted and unquoted strings may contain spaces. Strings are trimmed however. Names may also include colons, semicolons, carrets, tildes, commas, and in fact all printable characters from the lower ASCII set except ()[]:, . The NHX format tags have been taken from: http://www.genetics.wustl.edu/eddy/forester/NHX.html. The format also allows ordinary comments which are not removed. However, these cannot be combined with NHX tags for any particular node. The program allows more values for the &&NHX:D tag: Y=Duplication, N=Speciation, A=AncientDup, I=IntermediateDup, R=RecentDup, H=HGT, E=EndNode. A :R tag is added to the NHX standard allowing to insert a reference. A :OL tag is added to the NHX standard allowing to store the Orthology Level assigned by LOFT (output only). It is allowed to include spaces (and line feeds) within NHX tags (spaces are included within string tags - all strings are trimmed). To comply with ClustalW standards, it interpretes integer comments in internal nodes as bootstrap values. To comply with PhyML standards, internal nodenames that are integers are also interpreted as bootstrap values. Nodenames should be unique. Duplicate nodenames are extended with "_n_" tags, where n is a counter. A tree does not need to end with a semicolon ';'.