Homology

After identifying all domains for each SR protein a search inside the Superfam database provided by the NCBI (accessed 01/18/2009) for the SCOP ID of the annotated domain was done the following way:

1.

Download the NCBI Superfam database

2.

Extract all species which posses an annotated domain which is known to be a part of an SR protein

3.

Build a valuated species-based newick tree for each domain

4.

Build a valuated species-based newick tree for all domain combinations

The algorithmic procedure can be summarized like this:

Algorithm (exemplary):

1. Extract RRM with superfamily SCOP-ID: 54929

   awk 'BEGIN { FS = "[\t]" } ; { if($9 == "54929") { print }};' 964.ass.tab > 54929.tab
   

2. Count the number of species with a concrete RRM and print both species-id and number to domaincount_species.tab

   awk '{print $1"\t"$2}' 54929.tab | sort -u | awk '{print $1}' | uniq -c | awk '{print $2"\t"$1}' > domaincount_species_rrm.tab
   

3. Join ncbi_names_id.tab and domaincount_species_rrm.tab on <short code> to get a file with tab-seperated values like <(super-)fam id> <taxon name> <count of domains in taxon>

   map_domain_counts.sh ncbi_names_id.tab domaincount_species_rrm.tab > domaincount_species_and_names_rrm.tab
   

4. Extract the name and domain count from the domain count file (=strip ncbi id)

   awk 'BEGIN { FS = "[\t]" } ; {print $2"\t"$3}' domaincount_species_and_names_rrm.tab > genecount.tab
   

5. Build the newick tree via Axel Wintsches tax2 and addStats scripts

   awk 'BEGIN { FS = "[\t]" } ; {print $1}' domaincount_species_and_names_rrm.tab | sort -u > ncbi_id_all.tab
   tax2 -f ncbi_id_all.tab -n > basic_tree.tree
   addStats.pl genecount.tab basic_tree.tree > annotated_tree.tree
   

6. Plot the newick tree accordingly

   njplot.linux annotated_tree.tree
   

Files: assets/files.tar.gz