uAnalyze(SM) is a web-based tool for analyzing high-resolution melting data of PCR products. PCR product sequence is input by the user and recursive nearest neighbor thermodynamic calculations used to predict a melting curve similar to uMELT(http://www.dna.utah.edu/umelt/umelt.html). Unprocessed melting data are input directly from LightScanner-96, LS32, or HR-1 data files or via a generic format for other instruments. A fluorescence discriminator identifies low intensity samples to prevent analysis of data that cannot be adequately normalized. Temperature regions that define fluorescence background are initialized by prediction and optionally adjusted by the user. Background is removed either as an exponential or by linear baseline extrapolation. The precision or, “curve spread,” of experimental melting curves is quantified as the average of the maximum helicity difference of all curve pairs. Melting curve accuracy is quantified as the area or “2D offset” between the average experimental and predicted melting curves. Optional temperature overlay (temperature shifting) is provided to focus on curve shape. Using 14 amplicons of CYBB, the mean + / - standard deviation of the difference between experimental and predicted fluorescence at 50 percent helicity was 0:04 + / - 0:48°C. uAnalyze requires Flash, is not browser specific and can be accessed at http://www.dna.utah.edu/uv/uanalyze.html.

Primer3 is a widely used program for selection of oligonucleotide primers for PCR. The websites used for implementation of Primer3 have recently been updated. PCR requires Mg(2+)(,) which has a significant dsDNA stabilizing effect that must be taken into account when designing PCR primers. The data sets and formulas used to correct for salt concentrations have been updated in Primer3 to give better prediction of melting temperature (T(m)). The liberal combination of different formulas for monovalent and divalent salt correction can lead to different results, depending on the formula chosen by the user. Using published T(m) for 475 different oligonucleotides, it is shown that the combination of the implemented conversion of divalent to monovalent cation concentration works well with one salt correction formula but not with an alternative one. Use of a more recently described alternative formula would lead to equally good T(m) predictions if divalent cations are present. The proper selection of compatible primer pairs depends on the choice of a good combination of salt correction formulas. Currently the SantaLucia salt correction formula should be used if Mg(2+) is present. The alternative formula should be updated to its recent form for future releases.

UNLABELLED: uMelt(SM) is a flexible web-based tool for predicting DNA melting curves and denaturation profiles of PCR products. The user defines an amplicon sequence and chooses a set of thermodynamic and experimental parameters that include nearest neighbor stacking energies, loop entropy effects, cation (monovalent and Mg(++)) concentrations and a temperature range. Using an accelerated partition function algorithm along with chosen parameter values, uMelt interactively calculates and visualizes the mean helicity and the dissociation probability at each sequence position at temperatures within the temperature range. Predicted curves display the mean helicity as a function of temperature or as derivative plots. Predicted profiles display stability as a function of sequence position either as 50% helicity temperatures or as the helicity probability at specific temperatures. The loss of helicity associated with increasing temperature may be viewed dynamically to visualize domain formation within the molecule. Results from fluorescent high-resolution melting experiments match the number of predicted melting domains and their relative temperatures. However, the absolute melting temperatures vary with the selected thermodynamic parameters and current libraries do not account for the rapid melting rates and helix stabilizing dyes used in fluorescent melting experiments. uMelt provides a convenient platform for simulation and design of high-resolution melting assays. AVAILABILITY AND IMPLEMENTATION: The application was developed in Actionscript and can be found online at http://www.dna.utah.edu/umelt/umelt.html. Adobe Flash is required to run in all browsers.

To diagnose AIDS patients, doctors are starting to use the methodology of amplification induced by thermal annealing and by adding specific primers to imitate the polymerase chain reaction of the denatured DNA mutated by several types of human immunodeficiency viruses. An adaptive resonance theory (ART) neural network working in concert with an artificial intelligence (AI) rule-based system is shown to be efficient for enhancing the choice of the needed good primers. Because of the tradeoff between the simplicity of primers and the specifics of primers, doctors prefer to diagnose different viral types by administering patients with different sets of good primers. Thus, AI provides ART with both the background for self-organization and the foreground for the final goal.>

A neural network computing model for image sorting by means of invariant clustering is described. Such image sorting by clustering belongs to the class of inverse problems of image processing. One of the biomedical applications is to diagnose AIDS virus-mutated DNA by a powerful recombinant DNA technology, the polymerase chain reaction (PCR), based on judicious choice of oligonucleotide primers for the synthesis of an imprecisely known DNA sequence due to the unknown viral infections. Thus, a smaller set of primers among a larger open set, generated by a feature rule-based system, must be found. A neural network is needed to do nondestructive sorting and maintain the chemical bases (G, C, A, T) of the chosen primers. The problem is solved by an ART-1 neural network for feature extraction by clustering and an ART-4 (a hybrid ART-1) neural network for image sorting by clustering. The initial condition of ART-1 is precisely the complement of the initial condition of ART-4. The ART-1 whitens the cluster leader by the veto power of newcomers, while the ART-4 edifies and preserves the cluster leader by the consent of followers. A feature based expert system provides a powerful training environment for the ART-4 neural network to further discover new cluster rules