[Biojava-dev] AFPChain getSimilarity function

[Andreas Prlic]

Hi Ambikesh,

there is a large number of good papers about structure alignment out
there. A possible start is the wikipedia page for it and then dig
deeper in some of the articles that are referenced there.

http://en.wikipedia.org/wiki/Structural_alignment

about the scores:

getSimilarity gives the seq similarity in the alignment
getSimilarity1 and getSimilarity2 are the coverage relative to each of
the two proteins.

Andreas



On Thu, Jan 6, 2011 at 2:46 AM, Ambikesh Jayal <ambi1999 at gmail.com> wrote:
> Dear Andreas,
> Thanks. I will have a look a the functions you mentioned. What about the
> functions AFPChain.getSimilarity() getSimilarity1(), getSimilarity2()? I
> thought they were the main functions for finding similarity between the two
> structures.
> I am new to genetics and protein structure similarity measures. Do you know
> of any easy to understand paper that will give me a deeper insight into
> protein structure matching?
> Thanks
> Ambi.
>
> On Thu, Jan 6, 2011 at 4:15 AM, Andreas Prlic <andreas at sdsc.edu> wrote:
>>
>> Hi Ambikesh,
>>
>> Sorry for the lack of documentation... will add javadoc for this as
>> well as set a page up on the wiki.
>>
>> There are a few methods that provide access to various scores. If you
>> want to compare the similarity between the two structures,  probably
>> the most important values to look at are:
>>
>> getNrEQR()  the number of structurally equivalent residues
>>
>> getTotalRMSDOpt() the RMSD of the alignment
>>
>> getProbability() the Probability of the alignment (FATCAT) or the Z-score
>> (CE)
>>
>> getTMScore() returns the TM-score
>>
>>  Does that make more sense?
>>
>> Andreas
>>
>>
>>
>> On Wed, Jan 5, 2011 at 10:06 AM, Ambikesh Jayal <ambi1999 at gmail.com>
>> wrote:
>> > Hi,
>> > Is there a document that explains the purpose of different functions in
>> > the
>> > AFPChain class? Basically I am interested to find how similar two
>> > protein
>> > structures are. The following four functions seem to be returning values
>> > indicating similarity but I am not sure what is the main difference
>> > between
>> > them. So it will be helpful if someone can explain.
>> > org.biojava.bio.structure.align.model.AFPChain
>> > getSimilarity()
>> > getSimilarity1()
>> > getSimilarity2()
>> > getAlignScore()
>> > I ran the CE and FATCAT algorithms for two proteins and got the
>> > following
>> > results. Not sure how to interpret them. Any help?
>> > CE, name1 = "1cdg.A"; name2 = "1tim.B";
>> > **** afpChain.getSimilarity() 0.1951219512195122
>> > **** afpChain.getSimilarity1() 30
>> > **** afpChain.getSimilarity2() 83
>> > **** afpChain.getAlignScore()693.7852267498898
>> >
>> >
>> > FATCAT, name1 = "1cdg.A"; name2 = "1tim.B";
>> > **** afpChain.getSimilarity() 0.17647058823529413
>> > **** afpChain.getSimilarity1() 21
>> > **** afpChain.getSimilarity2() 58
>> > **** afpChain.getAlignScore()186.61939969257776
>> >
>> >
>> >
>> >
>> > FATCAT name1 = "1cdg.A"; name2 = "1cdg.A";
>> > **** afpChain.getSimilarity() 1.0
>> > **** afpChain.getSimilarity1() 100
>> > **** afpChain.getSimilarity2() 100
>> > **** afpChain.getAlignScore()2040.0
>> >
>> > Kind Regards,
>> > Ambi
>