NAME

AnalyseSeqs - Analyse a set of sequences of common length

SYNOPSIS

AnalyseSeqs [-X[bswn]] [-Q] [-M{mask}[+|!]] [-D{H|A|G}] [-d{S|H|D|B}]

DESCRIPTION

AnalyseSeqs reads a set of sequences from stdin and tries a variety of methods for sequence analysis on them. Currently available are:
Statistical geometry for quadruples of sequences; THIS IS PRELIMINARY AND NOT WELL TESTED BY NOW.
split decomposition; neighbour joining and Ward's variance method for reconstructing phylogenies using various distance measures. For statistical geometry and the cluster methods PostScript output is available.
The program continues reading until it encounters one of the separator characters '@' or '%'. Only sequences of alphabetical characters or of a specified alphabet are processed, all other lines are ignored. The program stops reading if it either encounters an EOF condition, or if there are no valid sequence data between two lines beginning with separator characters.
A list of taxa names can be specified in the input stream. The list begins with a line beginning with '*'. Optionally, a file name prefix [fn] for the PostScript output can be specified in this line. The entries have the form 'x : Taxon', where x is the number of taxon, i.e., of the corresponding entry in the list of input sequences. The taxa list need not be complete. It must end, however, with a line beginning with '*' or any of the separator characters. The taxa list is printed on top of the output. The specified taxa names are used as labels in the PostScript output.

OPTIONS

-X[bswn]: specifies the analysis methods to be used.
[b]: Statistical Geometry. A PostScript file named '[fn_]box.ps' giving a graphical representation of the statistical geometry is created. The resulting box is a good measure of 'tree likeness' of the data set. This is the default.
[s]: Split decomposition.
[w]: Cluster analysis using Ward's method. A PostScript file named '[fn_]wards.ps' is created containing a drawing of the tree.
[n]: Cluster analysis using Saitou's neighbour joining method. A PostScript file named '[fn_]nj.ps' is created containing a drawing of the tree.
-Q: indicates that a statistical geometry analysis is to be performed comparing four data sets, for instance to confirm the significance of a proposed phylogeny. This option is only useful for statistical geometry analysis and hence the -X option is ignored. Each of the four data sets must be of the form
* [filename_prefix]
# number
[list of taxa names]
*
list of sequences
%
where number is 1,2,3,4 for the four groups to be compared.
-M{mask}[+|!]: allows one to specify a mask for the input file. '{mask}' can be one of the following letters indicating a predefined alphabet or the %-sign followed by all characters to be accepted. A + sign at the very end of the mask indicates that the input is to be handled case sensitive. Default is conversion of the input to upper case. A ! sign can be used to convert the input data to RY code: GgAaXx -> R, UuCcKkTt -> Y, all other letters are converted to *.
-Ma: all letters A-Z and a-z.
-Mu: uppercase letters.
-Ml: lowercase letters.
-Mc: digits [0-9].
-Mn: all alphanumeric characters.
-MR: RNA alphabet (GCAUgcau).
-MD: DNA alphabet (GCATgcat).
-MA: Amino acids in one-letter code.
-MS: Secondary strcutures coded as '^.()'
-M%alphabet: use the specified alphabet.
-D: specifies the algorithm to be used for calculating the distance matrix of the input data set. Available are
-DH: Hamming Distance
-DA[,cost]: Simple alignment distance according to Needleman and Wunsch. A gap cost different from 1. can be specified after the comma.
-DG[,cost1,cost2]: Gotoh's distance with gap cost function g(k) = cost2+cost1*(k-1). cost2<=cost1 has to be fulfilled. Default values are cost1=1., cost2=1., yielding the same distance as option A.
ONLY THE HAMMING DISTANCE IS WELL TESTED BY NOW !!!
-d: specifies the edit cost matrix to be used. Available are
-dS: simple distance. Indel and substitution of different characters all have cost 1. The indel cost can be set by specifying the gap costs with the algorithm options -DA and -DG. This is the default.
-dH: A distance matrix for RNA secondary structures. Inspired by Hogeweg's similarity measure (J.Mol.Biol 1988). Gap-function is set automatically.
-dD: Dayhoff's matrix for amino acid distances.
-dB: Distinguish purines and pyrimidines only. CAUTION this option of course influences only the calculation of distances. It does NOT affect computation of the statistical geometry. This is done directly on the sequences. If you want to do statistical geometry on RY sequences use the ! sign with the -M option, for instance -MR!.

REFERENCES

The method of statistical geometry has been introduced by M. Eigen, R. Winkler-Oswatitsch and A.W.M. Dress (Proc Natl Acad Sci, 85:1988,5912). The method of split decomposition was proposed by H.J. Bandelt and A.W.M. Dress (Adv Math, 92:1992,47). The variance method for cluster analysis is due to H.J. Ward (J Amer Stat Ass, 58:1963,236). The neighbour joining method was published by Saitou and Nei (Mol Biol Evol, 4:1987,406).

This program is part of the Vienna RNA Package

WARNING

This is the beta test version. Some options or combinations of options may still produce nonsense. Please send bug reports to ivo@tbi.univie.ac.at.

VERSION

This man page is part of the Vienna RNA Package version 1.2.

AUTHOR

Peter F Stadler, Ivo L. Hofacker.

BUGS

Comments should be sent to ivo@itc.univie.ac.at.