A pentapeptide-based method for protein secondary structure prediction.

We present a new method for protein secondary structure prediction, based on the recognition of well-defined pentapeptides, in a large databank. Using a databank of 635 protein chains, we obtained a success rate of 68.6%.

Fast DNA sequencing.

A method for fast nucleotide sequencing is described. It is based in the selection of well-known small oligomers able to be hybridized with the unknown target. The selected oligomers are afterwards ordered following a simple statistical approach.

Biocryptography.

A new cryptology that uses biomolecules as carriers of hidden information is described here. The huge dimension of biopolymers permits the insertion there of private messages and the rapidity and specificity of biomolecular interactions facilitate the identification of classified communications. Otherwise, this process can be implemented in a virtual context.

Biochemical identification of prime numbers.

A biochemical technique is proposed whereby prime numbers may be identified.

Secondary structure of proteins and three-dimensional pattern recognition.

We describe a new, parameter-free, method to predict the secondary structure of a protein. It is based on the recognition of well-defined pentapeptides, which allows the discrimination of alpha helices, beta sheets and random coils in proteins. Presently, a success rate of about 65% is achieved for a three-state model.

On the determination of internal nodes of an evolutionary dendrogram.

In this communication the determination of intermediary nodes in evolutionary dendrograms of proteins is analyzed, based on the molecular mechanisms involved in these transitions and the probability of acceptance of the amino acid substitutions.

Neural system states and networks.

A neural circuit at the human cerebellar cortex and nuclei is interpreted in terms of an optimized code and a reliable network.

On the nature of mutations in molecular evolution.

In this work it is proposed tht in evolution amino acid substitutions implying strong physico chemical and structural differences are more revelant and more frequent that substitutions between similar amino acids. This analysis is made over a group of protein families representing about 10000 substitutions and as examples the evolutionary trees of fibrinopeptides A and calcitonins were constructed and compared.

An optimized interconnected biomolecular computer.

A biomolecular device for information-handling is described where the hardware is optimized and a neural network increases its reliability.

Neural network in the transcription-translation process in eukaryotic cells and viruses, a comparison.

The transcription-translation process in the virus-host cell system is analyzed in terms of a virtual neural network. In contrast to findings in the transcription-translation process in eukaryotic cells, an inverse correlation between the frequency of protein amino acids and code degeneracy and between code degeneracy and the rate of amino acid substitutions was not found. Misuse of the virtual host cell network by virus protein synthesis could be attributed to the different amino acid composition of the virus proteins (input unit) studied here.

Neural network in the transcription-translation process.

The transcription-translation process is represented by a virtual neural network. The input unit corresponds to the transcription process, the hidden unit to the mRNA-ribosome-tRNA interaction and the output unit to the final state of the protein. A correlation was found between the natural amino acid frequency and the degree of codon degeneracy and an inverse correlation between the code degeneracy and the number of amino acid substitutions in the protein.

Algorithm for assisting medical diagnosis.

An algorithm for medical diagnosis assistance is described. In this algorithm, the patient is identified following the simultaneous evaluation of its degree of similarity with a given disease, as well as the extent of disagreement with the rest of diseases considered in the differential diagnosis. In an example the algorithm is applied to the diagnosis assistance of some neuropathies.

Pattern recognition of the secondary structure of proteins (alpha-helix and beta-structure).

In this paper, an algorithm for the pattern recognition of secondary structure of proteins is proposed. The procedure simultaneously evaluates the contribution of all the residues of a given peptide to its conformation. By means of the algorithm it is possible to select from a universe of well known proteins the most representative alpha-helix and beta-structure peptides, and to use these peptides, as screening matrices to define the unknown structure of any peptide.

Phylogenetic tree of tRNAs using a simple algorithm.

A simple method for phylogenetic tree construction is described. In this method each node is calculated considering the distance between the elements and the difference between these elements and an average element, allowing the selection of the most probable node. Two examples of tRNA phylogenies (E.

Codon evolution and conservation of the reading phase in genetic code translation.

The description of the optimized evolution of a code based on 4 nucleotides involves a sequential transition of codons, formed firstly by monomers evolving to dimers and then to triplets, in accordance with the progressive increase of the number of amino acids to be coded. The successive increase in the size of these codons during evolution implies changes in the phase reading of the genetic message, which could become chaotic. In order to overcome this constraint, this paper proposes a codon evolution where two things occur simultaneously: codons change in size and there is an alternation of the molecule which holds the information.

Analogical representation of the genetic code in terms of the virtual base "e" and codons of triplets.

The analogical representation of the genetic code on base "e" is described. The possible use of the "e" base in codification processes is emphasized.

Algorithm for construction of phylogenetic trees.

An algorithm is described for the construction of phylogenetic trees. The algorithm is based on the progressive correction of data along the tree construction. For the correction, the average value of the difference between each pair of neighbour elements to the rest of the table is considered.

A general algorithm for assisting medical diagnosis.

An algorithm for assisting medical diagnosis is designed; this algorithm is highly reliable and can be given general use. The method is based on the comparison of the column vector of the symptoms of a given patient with each of the columns of disease symptoms which appear on a matrix. On this matrix, the symptoms have been arranged according to the hierarchic order of each of the symptoms, which facilitates the creation of a file.

A simple algorithm for assisting medical diagnosis.

This is a description of an algorithm to assist in the selection of the most probable diagnosis of a given patient. This algorithm is based on the comparison of a scaled column vector of the symptoms displayed by the patient, and of each column of symptoms appearing in a disease matrix. Moreover, the same algorithm gives the degree of similarity of the diseases compared by the differential diagnosis, as well as the hierarchic order of the symptoms that characterise any disease.

Conservation of the secondary structure of protein during evolution and the role of the genetic code.

In this communication we demonstrate, in a group of modern proteins, following an algorithm described by Argyle (1980), that the ordination of the amino acids in terms of the most frequent substitutions agrees with the conservation of the alpha-helix, beta-sheet, and beta-turn formation tendencies during evolution. The same correspondence has been demonstrated for the conservation of the physico-chemical properties in the amino acid substitutions. Both parameters are similar in showing higher correlation with the most frequent amino acid substitutions than with the feasibility of changes at the level of the respective codons.

Dissimilar rates in molecular evolution.

In this work we present an evolutionary tree based on the differences in the physico-chemical properties involved in amino acid substitutions, instead of considering, for its construction, only the number of changes between species. Phylogenetic trees were constructed from the differences in bulkiness, refractivity index, hydrophobicity, polarity and optical rotation of 9 vertebrate calcitonins. A correlation of the form y = a xb was found between the number of changes (x) and the differences in any given physico-chemical property (y).

Conservation of physico-chemical amino acid properties during the evolution of proteins.

Based on a similarity ring constructed from a substitution probability matrix, we have analyzed the conservation of some amino acid properties in the evolution of proteins. Refractive index and bulkiness are highly conserved, hydrophobicity and polarity are fairly well conserved while optical rotation appears to be a less relevant property. On the other hand, the analysis of the correspondence between phenotype and genotype shows that the most frequent amino acid substitutions in proteins do not always correspond to the most feasible codon changes.