Comparison of different melting temperature calculation methods for short DNA sequences Authors

Motivation: The overall performance of several molecular biology techniques involving DNA/DNA hybridization depends on the accurate prediction of the experimental value of a critical parameter: the melting temperature. To date, many computer software based on different methods and/or parameterizations are available for the theoretical estimation of the experimental melting temperature value of any given short oligonucleotide sequence. However, in most cases, large and significant differences of melting temperature estimations are obtained when using different methods. Thus, it is difficult to decide which melting temperature value should be trusted. In addition to this, it seems that most people use these methods being unaware of their limitations, which are well described in the literature but not stated properly or restricted on the input of most of the web servers and standalone software that implement them. Results: A quantitative comparison of the similarities and differences among some published DNA/DNA melting temperature calculation methods is reported. The comparison was carried out for a large set of short oligonucleotide sequences ranging from 16 to 30 nucleotides long and spanning the whole range of CG content. The results showed that significant differences are observed among all methods, which in some cases depend on oligonucleotide length and CG content in a non-trivial manner. Based on these results, regions of consensus and disagreement for the methods in oligonucleotide feature space are reported. Due to the lack of enough experimental data, a fair and complete assessment of accuracy for the different methods is not yet possible. Though this limitation, a consensus melting temperature with minimal error probability is calculated by averaging the values obtained from two or more methods that exhibit similar behavior at each particular combination of oligonucleotide length and CG content class. Using a total of 348 DNA sequences in a size range between 16 and 30 mers for which experimental melting data is available, we demonstrate that the consensus melting temperature is a robust and accurate measure. It is expected that the results of this work will constitute a helpful set of guidelines to follow for the successful experimental implementation of various molecular biology techniques such as quantitative PCR, multiplex PCR and the design of optimal DNA microarrays. Availability: A binary software distribution to calculate the consensus melting temperature described in this work for thousands of oligonucleotides simultaneously is freely available for the LINUX operative system upon request to the authors or from our web site …