Hi, I am trying to get some autocorrelation analysis based on positions of DNA words. Given a certain DNA motif or word, you can have its positions on the genome. I am trying to figure out if there is any regularity (pattern) in that. What I did is to determine pair-wise distances and see how many pairs have a certain distance. The result indicates that is more likely to have such words at a given relative distance than at others, but it is not so easy to interpret. Is there any way to test this? There is a lot of functions in R about testing autocorrelation on time-series data or 2D geographical data .. but these do not seem really appropriate here to me. Also, it would be nice to test weather two positions sets for two different motifs are correlated .. thanks for any suggestion! mattia

Hi, Mattia. I am not an expert in the field of nonlinear dynamics, deterministic chaos, fractals etc but I know a little bit about it. You should tka a look at the packages related to the chaotic (nonlinear) data series and related analytical tools like false nearest neighbors, average mutual information, Lyapunov exponent etc. There are few R packages that contain related information and applications. I must warn you that this area of mathematical science is very demanding (although very appealing also). If you are very keen to start the journey read the available basic texts before doing any experiments of yours. There are many applications of these tools in the field of molecular biology, and especially biochemistry. Hope this helps. Zeljko Debeljak, PhD Medical Biochemistry Specialist Osijek Clinical Hospital 2009/5/25, mattia pelizzola <mattia.pelizzola at="" gmail.com="">: > Hi, > > I am trying to get some autocorrelation analysis based on positions of > DNA words. Given a certain DNA motif or word, you can have its > positions on the genome. I am trying to figure out if there is any > regularity (pattern) in that. What I did is to determine pair-wise > distances and see how many pairs have a certain distance. The result > indicates that is more likely to have such words at a given relative > distance than at others, but it is not so easy to interpret. Is there > any way to test this? There is a lot of functions in R about testing > autocorrelation on time-series data or 2D geographical data .. but > these do not seem really appropriate here to me. > Also, it would be nice to test weather two positions sets for two > different motifs are correlated .. > > thanks for any suggestion! > > mattia > > _______________________________________________ > Bioconductor mailing list > Bioconductor at stat.math.ethz.ch > https://stat.ethz.ch/mailman/listinfo/bioconductor > Search the archives: > http://news.gmane.org/gmane.science.biology.informatics.conductor >