Application of expert networks for predicting proteins secondary structure.

The present study utilizes expert neural networks for the prediction of proteins secondary structure. We use three independent networks, one for each structure (alpha, beta and coil) as the first-level processing unit; decision upon the chosen structure for each residue is carried out by a second-level, post-processing unit, which utilizes the Chou and Fasman frequency values Falpha and Fbeta in order to strengthen and/or deplete the probability of the specific structure under investigation. The highest prediction case was 76%. Our method requires primitive computational means and a relatively small training set, while still been comparable to previous work. It is not meant to be an alternative to the determination of secondary structure by means of free energy minimization, integration of dynamic equations of motion or crystallography, which are expensive, time-consuming and complicated, but to provide additional constrains, which might be considered and incorporated into larger computing setups in order to reduce the initial search space for the above methods.

Electrical impedance scanning: a new approach to skin cancer diagnosis.

BACKGROUND/AIMS: Skin cancer diagnosis depends, to a great extent, on visual inspection and histopathological examination of excised tissues. The aim of this study is to evaluate the ability of electrical impedance scanning to differentiate between benign and malignant skin lesions. METHODS: A preclinical study was conducted on 40 nude mice injected subcutaneously with a human melanoma strain. Impedance measurements were recorded every week to correlate electrical changes with tumor growth and histological findings. A clinical study was also performed on 178 human suspicious skin lesions before excision. The impedance measurements were correlated to the histopathological results. RESULTS: Normalized conductivity and capacitance, recorded on growing skin tumors in nude mice, were shown to change relative to lesion size. Necrosis, present in most of the larger lesions, was associated with a decrease in the electrical conductivity. Similar electrical parameters were used to classify human melanoma lesions with 92% sensitivity and 67% specificity. In addition, four out of five BCC lesions were correctly diagnosed. Moreover, dysplastic lesions were diagnosed with 91% sensitivity and 59% specificity. For comparison, physicians diagnosed melanoma lesions with 75% sensitivity and 87% specificity and dysplastic lesions with 46% sensitivity and 80% specificity. CONCLUSIONS: The animal study showed that electrical impedance measurements reflect morphological changes related to the growth of a cancerous skin lesion. These findings are in agreement with a preliminary clinical study. Electrical Impedance Scanning can therefore be considered as an objective and non-invasive tool for differentiation between benign and malignant skin lesions.

Novel EIS postprocessing algorithm for breast cancer diagnosis.

A new postprocessing algorithm was developed for the diagnosis of breast cancer using electrical impedance scanning. This algorithm automatically recognizes bright focal spots in the conductivity map of the breast. Moreover, this algorithm discriminates between malignant and benign/normal tissues using two main predictors: phase at 5 kHz and crossover frequency, the frequency at which the imaginary part of the admittance is at its maximum. The thresholds for these predictors were adjusted using a learning group consisting of 83 carcinomas and 378 benign cases. In addition, the algorithm was verified on an independent test group including 87 carcinomas, 153 benign cases and 356 asymptomatic cases. Biopsy was used as gold standard for determining pathology in the symptomatic cases. A sensitivity of 84% and a specificity of 52% were obtained for the test group.