I did a blast search with a sequence of my protein of interest. I got around 4000 sequences in the result. How do I eliminate redundant sequences and multiple entries for the same organism from a blast search result? Using CD hit to set a cut off of 90% still, gives a huge number of redundant sequences. I have even gone down to 40% cutoff and still ended up with multiple of redundant entries. When I do sequence search in Pfam database, there also I get around 6000 sequences. But seed sequences are only 12. Which is the right way to obtain a good number of homolog sequences? thanks in advance...
Use NCBI syste. After BLASTing, use BLink, set display to remove identical entries and set the similarity value to something which gives you the desired number of sequences. If you have trouble, look up training systems on YouTube or ask your advisor.
Hi Lakshmeesha
If you are interested in a very large and diverse family in its entirety then there may be no alternative to working with large numbers of sequences. CD-HIT is a good way to remove redundancy, but an alternative approach which may be less work is to search in a redundancy-reduced database. Uniprot BLAST
http://www.uniprot.org/blast/
for example, can search in UniRef90 and UniRef50 which are redundancy-reduced to 90% and 50% sequence identity respectively. HHBlits
https://toolkit.tuebingen.mpg.de/#/tools/hhblits
searches in UniClust30.
Good luck
Daniel
Using CLANS will help to an extent. Feed the BLAST results to CLANS (if u want to use GUI, its available at MPI tool kit) and use the representative sequences of each CLAN for further analysis. Caution is required while playing with cutoff values to get more meaning full Clans. Manual crosscheck and curation is important after this step. Hope this will help. Best.
- Badges
- Science topic
- Similar topics
- Biological Science
- Molecular Biology
- Biomolecules
- Proteins