Phylogenetic Reconstruction Using an Unsupervised Growing Neural Network That Adopts the Topology of a Phylogenetic Tree

Abstract

We propose a new type of unsupervised, growing, self-organizing neural network that expands itself by following the taxonomic relationships that exist among the sequences being classified. The binary tree topology of this neutral network, contrary to other more classical neural network topologies, permits an efficient classification of sequences. The growing nature of this procedure allows to stop it at the desired taxonomic level without the necessity of waiting until a complete phylogenetic tree is produced. This novel approach presents a number of other interesting properties, such as a time for convergence which is, approximately, a lineal function of the number of sequences. Computer simulation and a real example show that the algorithm accurately finds the phylogenetic tree that relates the data. All this makes the neural network presented here an excellent tool for phylogenetic analysis of a large number of sequences.

Self‐organizing tree‐growing network for the classification of protein sequences

“…SOTA has been developed to classify protein/DNA sequences and construct a phylogenetic tree by a self-organizing tree-growing approach, a special type of Kohonen neural mapping (Dopazo & Carazo, 1997). In this work SOTA has been further developed in such a way that it can use directly aligned sequence as input (SOTA/SEQ) as well as a matrix of dipeptide composition (SOTA/ DP) and composition of other n-grams, such as AE12 and A2E4 (SOTA/n-gram).…”

“…We have used the self-organizing tree growing network (Dopazo & Carazo, 1997) to classify and reconstruct phylogenetic tree of protein families. The SOTA network is a special case of the unsupervised growing cell structure (GCS) (Fritzke, 1994), which in itself was derived from the Kohonen self-organizing map (Kohonen, 1990).…”

“…In general, the sequences to be used as input to a neural net program such as SOTA can be encoded in two different ways: ( I ) direct sequence encoding, which uses an indicator vector of binary numbers (0 or 1) to represent the "identity" of each residue in the sequence string (Casari et al, 1995;Andrade et al, 1997;Dopazo & Carazo, 1997). For a protein sequence, an amino acid is represented as a vector of 21 input units (20 zeros and a single one), which includes an extra unit for the gaps.…”

“…A detailed description of the learning algorithm of SOTA can be found in Dopazo and Carazo (1997). The following is a brief summary of the procedures (Fig.…”

Filtration of DNA Nucleotide Gene Expression Profiles in the Systems of Biological Objects Clustering

“…Simulation was carried out by software KNIME using SOTA algorithm clustering [10]. All meaningful principal components were taken for component analysis and the size of array of studied objects decreased to (96×94).…”