Topological maps of protein sequences

Abstract

A new method based on neural networks to cluster proteins into families is described. The network is trained with the Kohonen unsupervised learning algorithm, using matrix pattern representations of the protein sequences as inputs. The components (x, y) of these 20 x 20 matrix patterns are the normalized frequencies of all pairs xy of amino acids in each sequence. We investigate the influence of different learning parameters in the final topological maps obtained with a learning set of ten proteins belonging to three established families. In all cases, except in those where the synaptic vectors remains nearly unchanged during learning, the ten proteins are correctly classified into the expected families. The classification by the trained network of mutated or incomplete sequences of the learned proteins is also analysed. The neural network gives a correct classification for a sequence mutated in 21.5% +/- 7% of its amino acids and for fragments representing 7.5% +/- 3% of the original sequence. Similar results were obtained with a learning set of 32 proteins belonging to 15 families. These results show that a neural network can be trained following the Kohonen algorithm to obtain topological maps of protein sequences, where related proteins are finally associated to the same winner neuron or to neighboring ones, and that the trained network can be applied to rapidly classify new sequences. This approach opens new possibilities to find rapid and efficient algorithms to organize and search for homologies in the whole protein database.

Self‐organized neural maps of human protein sequences

“…In this section we summarize the standard formalism of the method that we have previously proposed (see Ferran & Ferrara [1991 for a detailed description).…”

“…Although network training is time consuming, once the topological map is obtained, the classification of a new protein is very fast. We have tested the method by considering both small (-10 sequences) and large (-450 sequences) learning sets of well-defined protein families (Ferran & Ferrara, 1991, 1992a. For small learning sets, we have also shown that the trained network is able to classify correctly mutated or incomplete sequences of the learned proteins (Ferran & Ferrara, 1991).…”

“…This approach should not be limited by database size, because the number of macromolecular families is expected to grow more slowly than the number of sequences. Recently, 2 different neural-network-based methods following this approach have been proposed (Ferran & Ferrara, 1991;Wu et al, 1992).…”

“…This approach has been applied to detect signal peptide coding regions (Arrigo et al, 1991) and to cluster small organic molecules of analogue structure into families of similar activity (Rose et al, 1991). We have proposed a method based on Kohonen's algorithm to cluster protein sequences into families according to their degree of sequence similarity (Ferran & Ferrara, 1991, 1992a. The network was trained using, as inputs, matrix patterns of 20 X 20 components derived from the dipeptide composition of the protein sequences.…”

“…We have tested the method by considering both small (-10 sequences) and large (-450 sequences) learning sets of well-defined protein families (Ferran & Ferrara, 1991, 1992a. For small learning sets, we have also shown that the trained network is able to classify correctly mutated or incomplete sequences of the learned proteins (Ferran & Ferrara, 1991). We have also found, using a learning set of 76 cytochrome c sequences belonging to different species, that the time evolution of the map during learning roughly resembles the phylogenetic classification of the involved species (Ferran & Ferrara, 1992b).…”