Question

What Distance Metric Is Appropriate For Clustering Thousands Of Dna Structures

2

Entering edit mode

12.7 years ago

grosy ▴ 100

Dear Biostars!

My aim is to clustering thousands of DNA structures

Thanks ahead of time!

clustering dna • 2.3k views

ADD COMMENT • link updated 11.7 years ago by Istvan Albert 102k • written 12.7 years ago by grosy ▴ 100

score 1 · Answer 1 · 2012-05-04

1

Entering edit mode

12.7 years ago

João Rodrigues ★ 2.5k

I would recommend Profit. You select which atoms you are calculating the alignment on, so for DNA I'd use C4 and P for example. Then use the ALLVSALL command to fit all of them. Remember that you don't need all vs all, because RMSD is symmetric ( rmsd(ab) == rmsd(ba) ) so this might save you some time.

ADD COMMENT • link 12.7 years ago by João Rodrigues ★ 2.5k

0

Entering edit mode

Thanks @Joao... but when i ran the program, it generated a matrix with error written on it.... Could please tell me why i m encountering the error matrix.

ADD REPLY • link 12.6 years ago by grosy ▴ 100

1

Entering edit mode

Matrix error? Are you sure you got the right Profit software? There is one, part of EMBOSS, that is for sequence analysis and not for structural alignment!

Otherwise, what is the error that you get exactly.

Clarification:

I would use something simple to begin with:

REFE dna_1.pdb
MOBI dna_2.pdb
ATOMS C4, P
FIT

Test on a couple of structures first. You can run Profit interactively just by typing 'profit' in your terminal window (assuming it's in your path).

I got a notification that you replied but I don't see the post anywhere.. nevertheless, RATOMS/RZONE are used to calculate the RMS over a specific atom set/zone after a previous fitting.

ADD REPLY • link 12.6 years ago by João Rodrigues ★ 2.5k

0

Entering edit mode

THANKS A LOT @Joao :-)) i gotit ......

ADD REPLY • link 12.6 years ago by grosy ▴ 100

0

Entering edit mode

@joao.. it doesn't accept more than 1000 structures :( and i have total of 4000 DNA structures

ADD REPLY • link 12.5 years ago by grosy ▴ 100

0

Entering edit mode

Download it and use it locally

ADD REPLY • link 12.5 years ago by João Rodrigues ★ 2.5k

0

Entering edit mode

i am sorry @joao to trouble.. but ya i have downloaded the software and trying to use it locally, when ever i am giving all the structures, its showing Error ==> Maximum structure count (1000) exceeded. Increase MAXSTRUC skipped structure

ADD REPLY • link 12.5 years ago by grosy ▴ 100

1

Entering edit mode

Go to the source code directory, edit the Profit.h file. There is a line starting with MAXSTRUC. Then recompile Profit and you should be good to go!