Bioinformatics Advance Access originally published online on June 9, 2006
Bioinformatics 2006 22(15):1917-1918; doi:10.1093/bioinformatics/btl277
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
tuple_plot: Fast pairwise nucleotide sequence comparison with noise suppression
Genome Analysis, Leibniz Institute for Age Research, Fritz Lipmann Institute Beutenbergstr. 11, 07745 Jena, Germany
*To whom correspondence should be addressed.
 |
ABSTRACT |
|---|
 TOP ABSTRACT 1 INTRODUCTION2 ALGORITHM AND IMPLEMENTATION3 RESULTSREFERENCES |
|---|
Summary: The program tuple_plot identifies and visualizes local similarities between two genomic sequences, typically 100 kb or longer, by applying the well-known dotplot principle. A dictionary of sequence words built from the input sequences serves to construct a task-specific expectancy model that is used to attribute significance values to pairwise word hits. The dictionary-based approach allows fast computation, the computation time scaling to O(N log N), depending on the size of the input sequences. The proposed scoring scheme appreciably increases the signal-to-noise ratio and may help to improve other word-based sequence comparison approaches.
Availability: tuple_plot is available at http://genome.fli-leibniz.de/software.html and may be used under GNU public license.
Contact: szafrans@fli-leibniz.de
|
1 INTRODUCTION |
|---|
|
TOPABSTRACT1 INTRODUCTION2 ALGORITHM AND IMPLEMENTATION3 RESULTSREFERENCES |
|---|
Dotplot analysis has been used for several decades as a technique for first-pass pairwise sequence comparison and its visualization (Fitch, 1969; Gibbs and McIntyre, 1970). It continues to be valuable since it principally allows for fast computations and is independent of assumptions that typically underlie alignment algorithms. Originally, a dot was drawn into a two-dimensional representation of the comparison space for any matching pair of sequence residues. In order to avoid an excess of signal in comparison with larger sequences it has proven useful to represent rectangular sections of the comparison space by single image pixels and blacken these pixels if the number of matching residues exceeds some threshold value (Staden, 1982). The program DOTTER allows to vary such a stringency cutoff through an interactive front end (Sonnhammer and Durbin, 1995). Another strategy to reduce noise is to require consecutive matches, i.e. matches of sequence words (tuples) rather than single residues, between the two sequences (Maizel and Lenk, 1981).
In practice, dot-matrix based sequence comparison often yields undesirably low signal-to-noise ratios, especially if the input sequences are very large (>10 kb) and only moderately similar. Here, we define ‘noise’ as the portion of hit signal (local similarity) which occurs independently of any homology relation between the sequences. We followed the intuition that noise in word-based sequence comparisons mostly arises from overabundant words. By applying a background hit model, implemented in a novel dot-matrix comparison tool tuple_plot, we show that dotplot sequence comparisons can produce results with a significantly reduced noise level. A convenient aspect of our solution is that the background model is derived from the input sequences alone.
|
2 ALGORITHM AND IMPLEMENTATION |
|---|
|
TOPABSTRACT1 INTRODUCTION2 ALGORITHM AND IMPLEMENTATION3 RESULTSREFERENCES |
|---|
tuple_plot performs nucleotide sequence comparisons in three steps: (1) preprocessing and parametrization of the comparison task, (2) analysis of the sequences and (3) the actual dot-matrix translation procedure. Initially, the program finds suitable parameters for the comparison, depending on the input sequences and program options. An adequate word size w is chosen such that approximately one random occurrence of any word ti is expected throughout the two input sequences S1 and S2, with lengths l1 and l2, respectively.
 | (1) |
 | (2) |
In order to obtain accurate background estimates for the rectangular comparison space we analyze the sequence composition of both input sequences. The diagnostic significance of observed word matches is inferred as follows: ni,1 and ni,2 occurrences of word ti found in the sequences will automatically result in ni,1 x ni,2 word hits. But, at most min(ni,1,ni,2) hits originate from a homology relation between the two sequences. The proposed scoring scheme uses these measures to attribute a score value to any observed hit for tuple ti.
 | (3) |
The scores are applied in the third computation step, the construction of the dot matrix.
The overall computational effort is approximately O(l w) for the construction of the tuple dictionary, O(4w) proportional to O(l) for the background correction and O(l2 w) for the computation of the dot matrix. With w being roughly proportional to log l the upper bound of the overall computational time scales to O(l log l). The program tuple_plot is written in C++ and has been extensively tested under the operating systems Solaris and Linux. It should work under any system which provides an ANSI-conforming C++ compiler. The command line interface of tuple_plot provides various options that influence the sequence comparison or the display of its results. The results are returned in PNG and HTML format. Forward and reverse matches are distinguishable by separate colors used for plotting.
We note that the proposed strategy for noise suppression can be applied to several other word-based sequence comparison approaches, like BlastZ (Schwartz et al., 2003), SSAHA (Ning et al., 2001), MUMmer (Delcher et al., 2002) or LAGAN (Brudno et al., 2003). We anticipate that it will help to shorten computation times and shift reported sequence similarities towards actual sequence homologies.
|
3 RESULTS |
|---|
|
TOPABSTRACT1 INTRODUCTION2 ALGORITHM AND IMPLEMENTATION3 RESULTSREFERENCES |
|---|
Initial trials with a naïve dotplot approach applied to large-size sequence comparison (BACs with insert sizes of 140 and 200 kb) showed that a tremendous portion of noise arises from microsatellites and other repetitive sequence elements (Fig. 1A). With the proposed suppression technique, noise signal is reduced to an acceptable amount and shows a homogeneous distribution (Fig. 1B). As shown, the increased specificity of the dotplot approach allows to discern similarities between intergenic sequences, well above noise level. Within a few CPU seconds, tuple_plot completes a 140 x 200 kb comparison of BAC sequences. tuple_plot has been tested on a variety of input sequences and, using a range of parameters, it produces highly robust reports on sequence similarities.
|
|
Acknowledgments |
|---|
The authors are indebted to U. Gausmann, G. Glöckner and K. Huse for testing developmental versions of tuple_plot. The authors also thank J. Sühnel and R. Sinha for helpful discussions. This work was supported by the German Ministry of Education and Research, grant 0312704E.
Conflict of Interest: none declared.
|
FOOTNOTES |
|---|
Associate Editor: Martin Bishop
Received on March 14, 2006; revised on May 19, 2006; accepted on May 29, 2006
|
REFERENCES |
|---|
|
TOPABSTRACT1 INTRODUCTION2 ALGORITHM AND IMPLEMENTATION3 RESULTSREFERENCES |
|---|
Brudno, M., et al. (2003) LAGAN and multi-LAGAN: efficient tools for large-scale multiple alignment of genomic DNA. Genome Res, . 13, 721–731
Delcher, A.L. (2002) Fast algorithms for large-scale genome alignment and comparison. Nucleic Acids Res, . 30, 2478–2483
Fitch, W.M. (1969) Locating gaps in amino acid sequences to optimize the homology between two proteins. Biochem. Genet, . 3, 99–108[CrossRef][Web of Science][Medline].
Gibbs, A.J. and McIntyre, G. (1970) The diagram, a method for comparing sequences. Its use with amino acid and nucleotide sequences. Eur. J. Biochem, . 16, 1–11[Web of Science][Medline].
Maizel, J.V.J. and Lenk, R.P. (1981) Enhanced graphic matrix analysis of nucleic acid and protein sequences. Proc. Natl Acad. Sci. USA, 78, 7665–7669
Ning, Z., et al. (2001) SSAHA: a fast search method for large DNA databases. Genome Res, . 11, 1725–1729
Schwartz, S., et al. (2003) Human-mouse alignments with BLASTZ. Genome Res, . 13, 103–107
Sonnhammer, E.L. and Durbin, R. (1995) A dot-matrix program with dynamic threshold control suited for genomic DNA and protein sequence analysis. Gene, 167, GC1–GC10[CrossRef][Web of Science][Medline].
Staden, R. (1982) An interactive graphics program for comparing and aligning nucleic acid and amino acid sequences. Nucleic Acids Res, . 10, 2951–2961
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  J. Krumsiek, R. Arnold, and T. Rattei Gepard: a rapid and sensitive tool for creating dotplots on genome scale Bioinformatics, April 15, 2007; 23(8): 1026 - 1028. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||