Lists all genes of all datasets with their unique sequence number and the labels read from the FASTA files.

The groups of orthologs are given as one per row, starting with a unique group identifier, followed by all group members, all separated by tabs.

More compact version of OrthologousGroups.txt . The groups of orthologs are given as matrix with group per row and one genome per tab-separated column. Numbers refer to entry number as listed in the file Map-SeqNum-ID.txt .

The OMA groups of orthologs stored in orthoxml format.

Each OMA group is provided as a separate Fasta file, with the species name as identifer. This format is particularly suitable as starting point for a phylogenetic reconstruction.

The textfiles in Output/PairwiseOrthologs are named according to {genome a}-{genome b}.txt and consist of a list of pairwise orthologs for the two given genomes. Every pair is listed only once, and in no particular order. Each line in the file contains one pair; all fields are separated by tabs. In the first two field, the unique IDs of the proteins are given. The next two fields contain the labels of the proteins, and in the last two fields, the type of orthology and (if any) the OMA group is given.

The pairwise orthologs stored in orthoxml format. Each group in the file will have orthologs between genes from only two genomes.

The hierarchical groups of orthologs in OrthoXML format. A detailed description of how these groups are computed is forthcoming.

The inferred species tree on which the hierarchical groups inference procedure is based, in Newick format.

For each toplevel HOG group we provide a separate Fasta file with all protein sequences clustered to it. This format is particularly suitable as starting point for a phylogenetic reconstruction of a gene tree.

This file is kept as a reference and contains all parameters used in the current run.

All hits found by ESPRIT are listed in this file. It is a list of contigs, ordered according to their position relative to the putative ortholog. Each line describes one contig, the fields are separated by tabs. In the first field, the fragment pair ID is printed; the next two fields contain the labels of the first and second fragments found in this hit. The forth and fifth fields contain the label of the corresponding full gene and its genome name. Then follows the distance difference between the two fragments and the number of positions between them (i.e. the gap); at last, an array is listed containing the IDs of all s3 genes corresponding to this hit.

ESPRIT often detects more candidate pairs than it will list in the hits.txt file, but not all of them survive the quality check. Still, if you want to see which triplets have been filtered out, have a look at dubious.txt where they are still listed. The file format is the same as for hits.txt .

If you want to recompute everything from scratch everytime the script is run, set this to false .

In the all-against-all phase, each genome pair is split in smaller junks of AlignBatchSize protein comparisons. The larger this number, the longer each unit runs, and the fewer files get produced. This allows to adjust the frequency of milestone steps (e.g. in case of computer crash) or to process few but large genomes with many CPUs efficiently.

Alignments which have a score lower than MinScore will not be considered. The scores are in Gonnet PAM matrices units.

Length tolerance ratio. If the length of the effective alignment is less than LengthTol * min( length(s1), length(s2) ), then the alignment is not considered.

During the stable pair formation, if a pair has a distance provable higher than another pair (i.e. StablePairTol standard deviations away) then it is discarded.

Tolerance parameter for the detection of differential gene losses using a third genome. The larger the tolerance, the liberaler the algorithm assigns orthologous relations. A detailed description is provided here.

Any sequence which is less than MinSeqLen amino acids long in regular genomes is not considered.

Enables/disables the generation of stable identifiers for OMA groups (and Hierarchical Groups if their computation enabled). The identifier consists of a prefix to determine the type of the group ('OMA' or 'HOG'), and a subsequence of the amino acid sequence uniquely present in this group. The computation of these ids might require a substantial amount of time. The ids are stored in the OrthoXML files only.

Enables/disables the computation of the hierarchical groups. This steps requires substantial computating power.

Define maximum amount of time (in sec) spent by the program for breaking every connected component of the orthology graph at its weakest link on a given taxonomic level. If set to a negative value, no timelimit is enforced.

The hierarchical groups require a (partially) resolved species phylogeny. With the parameter SpeciesTree the user can specify a phylogeny in Newick-format, or, by setting the variable to "estimate", compute a species tree based on the OMA Groups and use this one.

The cutoff of "average reachability within two steps" defines up to what point a cluster is split into sub-clusters.

Switches to disable the generation of certain output files if set to false . For very big analyses, disabling the generation of unused output might save a substantial amount of computing time and might drastically reduce the number of produced files.

You can either set this to true , which will enable ESPRIT and shut down the parts of OMA that are not directly needed for ESPRIT, or set it to false to make no use of ESPRIT at all.

Confidence level variable for contigs. This is the parameter tol described in the paper.

Minimal proportion of genomes with which contigs form many:1 BestMatches to consider that we might be dealing with fragments of the same gene. This is the parameter MinRefGenomes described in the paper, normalized by the total number of reference genomes.

Maximum overlap between fragments of same gene from different contigs.

Any sequence which is less than MinSeqLenContig amino acids long in contigs is not considered.

Minimum best score for BestMatch in scaffold recognition.