#############################################
UPDATED
Now consider following two proteins;
http://www.ncbi.nlm.nih.gov/protein/194443845
http://www.ncbi.nlm.nih.gov/protein/194443076
In previous case only the protein family name was mentioned. By reading in detail, we came to know that both proteins were 'different' subtypes of PBP. But here protein name, subunit, specie and strain name, explanation in CDS, COG, PFAM etc are same, only GI and locus is different. When I performed BLASTp of both sequences, I got following result;
Query cover = 94%
E value = 5e-32
Identity = 29%
By looking at these results, I can conclude that both of these proteins are different, but how can I prove it? since both of them have many similar features.
Need critical explanation to it.
#############################################
Consider following two proteins;
http://www.ncbi.nlm.nih.gov/protein/16764014
http://www.ncbi.nlm.nih.gov/protein/16766596
Both have same name (D-alanyl-D-alanine carboxypeptidase) and belong to same specie (Salmonella enterica subsp. enterica serovar Typhimurium str. LT2). But when I performed BLASTp among them, they shared only 42% identity in just 7% of the entire length. So should I consider them same or different ?
Their different lengths tell us from the start that they're not exactly the same thing. Looking further at the CDS information, we see that these are indeed annotated to be two different products, coded by different sequences. Comparing the relevant sections from the links you provided, it would appear that these two products appear on opposite strands of the DNA, from the /coded_by="" part in CDS. You might want to find the DNA region that codes these two and compare them instead, to see if that explains why they both have the same name
Be sure to look at the complete entry on the NCBI page -- each section might be important.
Just to be sure, I did my own alignment with clustalW on these; there's some conservation, but they are, for sure, different.
=======
http://www.ncbi.nlm.nih.gov/protein/16766596
...
Protein 1..477
/product="D-alanyl-D-alanine carboxypeptidase"
/EC_number="3.4.16.4"
/calculated_mol_wt=51826
...
CDS 1..477
/coded_by="NC_003197.1:3468654..3470087"
/note="'''penicillin binding protein 4; penicillin
sensitive; catalyzes the formation of D-alanine from
D-alanyl-D-alanine; one of four, DD-carboxypeptidase
low-molecular weight penicillin-binding proteins that
remove terminal D-alanine from pentapeptide side
chains'''"
http://www.ncbi.nlm.nih.gov/protein/16764014
...
Protein 1..403
/product="D-alanyl-D-alanine carboxypeptidase"
/EC_number="3.4.16.4"
/calculated_mol_wt=44326
...
CDS 1..403
/coded_by="complement(NC_003197.1:698401..699612)"
/note="penicillin-binding protein 5; removes C-terminal
D-alanyl residues from sugar-peptide cell wall precursors"
/transl_table=11
I think that their gene definition might be based on their own motifs. You should check the genomic coordinate and the similarity of DNA sequences (much different from a.a. similarity). There are many possibilities; alternative transcript, gene family, pseudogene.
However, in my initial opinion, given your analysis that there isn't significant relationship between them.
According to information provided by P.R.Krishna Behra they seem to be different proteins. Do you know if they are subunits of the same enzyme? In plants, for instance, the enzyme responsible for carbon fixation (Ribulose 1,5-bisphosphate carboxylase or Rubisco) consists of two subunits which have nothing to do in sequence, but of course they belong to the same enzyme and have the same name. Just letters S and L are added to make the distinction between the small and large subunit. The same happens to any heteromeric enzyme, like ATPases, for example.
From the above mentioned fruitful comments by Thomas Ingalls, Hong-Seok Ha, Krishna Behra and Gustavo Acevedo-Hernandez, it is now confirm that both of these proteins are different but belong to same class of protein family i.e. D-alanyl-D-alanine carboxypeptidase (a.k.a. penicillin-binding proteins).
But I still had a point here, that there should be a difference in name indicating its sub-class, as protein 16764014 is PBP-5, while 16766596 is PBP-4. Computational biologists works on thousands of proteins at a time, so it is not appropriate for us to report two different proteins from same specie having exact same name. I came across this case by chance, there might be few more. If I started checking thousands of proteins one by one, then what is the need of scripting or algorithm designing in data mining ?
Organisms often have multiple copies of paralogous genes, often because these genes are critical in some way. Thus, by definition the genes will have similar names and functions etc. They will have different genome locations and different IDs.
The key issue is to not identify any sequence by a name. You can only depend on a unique ID. Be it ensembl, uniprot, genbank, refseq or whatever. You can use the name as additional - human readable - information, but it is not an identifier.
This is somewhat complicated in newly identified genomes or strains as the identification of genes is fluid and unreliable.
@Christian Cole: I admit that these are paralogs, but there should be a naming scheme to differentiate between the paralogs, that is my point.
- Badges
- Science topic
- Similar topics
- Bioinformatics
- Bioinformatics and Computational Biology