INPARANOID: Eukaryotic Ortholog Groups
Release 7.0, June 2009


1. Introduction

InParanoid is a program for automatic identification of orthologs
while differentiating between inparalogs and outparalogs. An
InParanoid cluster is seeded by a reciprocally bestmatching ortholog
pair, around which inparalogs are gathered independently, while
outparalogs are excluded.  The InParanoid database is a collection of
pairwise ortholog groups aiming to include all 'completely sequenced'
eukaryotic genomes. By this we mean above 6X coverage, and less than
1% X letters in the protein sequences.


2. Online access

The InParanoid eukaryotic ortholog database is available both for
direct online access as well as for downloading. Online the user has
the option to view all clusters between two species, search for
clusters based on gene ID or free text search, as well as doing a
blast search based on a sequence. Online access is available at:
http://inparanoid.sbc.su.se.


3. Downloadable content

The current database is available for download at:
http://inparanoid.sbc.su.se/download/current
Previous versions can be found at:
http://inparanoid.sbc.su.se/download/old_versions

Analysed sequences:
Both the original and the processed (i.e. non-redundant, keeping only
the longest transcript) sequences used for analysis can be downloaded
either as fasta files or in XML format. In the XML directory, there is
also an XML schema file available called inparanoid-input.xsd. To
validate an XML sequence file, run for example xmllint with the
following command: 

xmllint --noout --schema inparanoid-input.xsd fileToValidate.xml


Ortholog clusters:
The ortholog clusters for all pairwise species comparisons are
available as orthoXML, SQL, HTML and raw text files (directory
table_stats). For more information about orthoXML see: 
http://www.orthoXML.org (coming soon). Included with the orthoXML 
files is a schema file called orthoXML.xsd that can be used to 
validate the orthoXML files using xmllint with the follwoing command: 

xmllint --noout --schema orthoXML.xsd fileToValidate.orthoXML


4. Database statistics

Version	Date	Species		Species_pairs	Ortholog_groups	Proteins_processed	Orthologous_proteins
2.0	05/03	7		21		57611		165186			86300
3.0	08/04	17		136		559269		303771			236979
4.0	04/05	26		325				463242
5.0	09/06	26		325		1501438		511758			368591
5.1	01/07	26		325		1501438		509483			405433
6.0	09/07	35		595		2642187		610047			501566
6.1	04/08	35		595		2642187		610047			501566
7.0	06/09	100		4950		42756512	1687023			1314592


5. Algorithmic differences compared to InParanoid 6

For release 7 we have used the program version 4.0.  The differences
compared to release 6 are:

- Tightened overlap criteria - all matches must now cover >50%.
- New Blast version (2.2.18) with compositional score matrix
adjustment was used; a second pass without low complexity filter was
used to get more reliable alignments.
- Lower score threshold (50->40 bits) was used as low complexity is
less of an issue.

Effects of algorithmic differences:
- In a few cases, orthologs that were previously found by InParanoid may be missed
in release 7 due to the stricter overlap criteria. On the other hand, low-scoring 
matches between 40 and 50 bits are now used, generating orthologs that were 
previously not detected.
- Much fewer low-complexity matches are included.  This mostly affect
the number of inparalogs in strongly biased composition proteomes, and
not so much the number of ortholog groups.

Other reasons exist why previous orthologs may be missing in the new
release.  In particular, new longer alternative transcripts can lead
to a too small overlap.


6. Stand alone program

InParanoid Version 4.0 will soon be published and be made available
upon request - email inparanoid@sbc.su.se to obtain a copy.