Thanks, both ideas are great. I did the first analysis before. It gives the evolutionary history of different motifs which are very interesting. About the second idea, if I use '-' substitute the motifs part, will it give the same result comparing to removing motifs?
How Can I Detect The Similarity Of Proteins According To Special Motifs In Them?
Hi, all. Thank you for reading this question.
I have some special sequences in proteins showed as follows:
All proteins contain one similar segments, and repeat in different times.
Take 'seq1' as an example, it has a sequence(not motif in biology meanings) 'EETELICLSDVTATSGAMEHSEVILKEREE' repeated 5 times. The number after colon is the position of this segment in protein.
seq1 EETELICLSDVTATSGAMEHSEVILKEREE:420# EETELICLNDVTSPLRAVEHSAVLLKDKVE:473# EETELICVNDVTSTSRTMGHSSVVLKENEE:527# EETELICLNDVTSPSRAMEHSTVFIEEKEE:580# EETELICLNDVTSTSEVAETPEDVLEGIEL:633
seq2 EETELICLNDVTSPLRAVEHSAVLLKDKVE:473# EETELICLNDVTSPSRAMEHSTVFIEEKEE:580
seq3 EETELICLNDFTSTSRAMEHSEVILKEREE:421# EETELICVNDVTSTSHAMEHSAVILKENEE:527# EETKLICLNEVTSTSRAMEHSAVVIEDKAE:580
seq4 EETELICLNDVTSPLRAVEHSAVLLKDKVE:473# EETELICLNDVTSPSRAMEHSTVFIEEKEE:580
seq5 EETELICLNDVTSPLRAVEHSAVLLKDKVE:473
I want to compute the similarity score between any two sequences based on mentioned sequence information. How can I do it? I want to consider two factors, the similarity of motifs and repeat times.
Thank you!
Before I joined them together as one sequence and used multiple alignment software get the similarity score of them. But I do not know if it is reasonable.
4 answers
One thing that can be done is to build a multiple alignment of the individual motifs: 1.1, 1.2, etc (where 1 indicates the source protein (seq1) and .2 is the second iteration of the motif). Then, use that alignment to look at a dendrogram (displaying the degree of difference in the motifs in visual form) or take it further with a phylogenetic tree to show (likely) evolutionary history. This would address one part of your objective - to show the relatedness of the motifs separate from the entire proteins.
Similarly, you can look at these five proteins with the motifs removed in order to get at their similarity outside the motif regions. That would actually be quite interesting.
• 3 views
•
link
written
by
Larry_Parnell
That cannot be predicted. You'll have to perform the analysis and compare results. That does not seem overly difficult with 5 sequences, plus perhaps an outlier, if it exists, that naturally does not contain the motif.
• 0 views
•
link
written
by
Larry_Parnell
Take a look at this suite of tools for protein motif analysis: http://bioware.ucd.ie/~compass/
There are several tools for comparing motif lists (Comparimotif) and discovering specific motifs in a set of sequences that accounts for convergent evolution (SLiMFinder).
• 0 views
•
link
written
by
Kieren Lythgow •
Some of these tools are for "amino acid segment" search in protein sequences like Logol, Stan or grappe. But if you just want string pattern recognition look here : http://www.genouest.org/spip.php?rubrique51
Have you consider Pattern Hit Initiated BLAST (PHI-BLAST) ? It was part of old versions of BLAST, not sure if it is retained in BLAST+.
Also check SCANMOT, a tool developed in my graduate lab for acheiving similar tasks.
• 0 views
•
link
written
by
Khader Shameer
Log in to answer this question.