Brief description of the dataset Thousands of short DNA sequences were generated in a computer, and the melting temperature of each of them was calculated by five different approaches. Several comparative measures were used to assess the differences and similarities of the observed melting temperatures. The approaches included the basic melting temperature calculation (denominated bas), the salt adjusted melting temperature calculation (denominated sal), and the nearest neighbors thermodynamic method based on the three most commonly used parameter sets. These thermodynamic parameter sets included the Breslauer table (Breslauer et al., 1986) denominated here as Th1 (Nearest-neighbor thermodynamic 1), the Santalucia table (Santalucia et al., 1996) called Th2 (Nearest-neighbor thermodynamic 2), and the Sugimoto table (Sugimoto et al., 1996) as Th3 (Nearest-neighbor thermodynamic 3). The mathematical expressions used to calculate the melting temperatures are described in detail here. The length of the DNA sequences was limited between 16 and 30 nucleotides, which is the most commonly used length range for PCR primer design and in situ synthetized oligonucleotide microarray. For each length, ten different CG-content classes were defined ranging between 0 and 100, thus covering the complete CG-content range. Finally, a total of 2.000 DNA sequences were randomly generated for each particular combination of length and CG-content class. For each sequence, the melting temperatures were calculated using the methods already described and several comparisons were carried out (see below). Comparative measures Several measures of similarity between the Tm values reported for any two methods were used in this work. All comparisons were done within a fixed oligonucleotide length and percentage of CG-content class, each involving a total of 2.000 oligonucleotide sequences. The calculated measures include the maximal observed absolute difference (MaxAD), the minimal observed absolute difference (MinAD), the average absolute difference (AveAD), the standard deviation of absolute differences (DevAD), the correlation coefficient (CC), and the percentage of cases where the absolute difference between calculated Tm values was equal or less than 5 (Per5C) and 3 (Per3C) Celcius degrees. The correlation coefficient was only calculated between thermodynamic methods because the other two methods have a variance equal to zero for oligonucleotides of fixed length and CG-content (they depend only on these two variables, which were of course identical at each grid point where the analysis was carried out). Breslauer,K.J., Frank,R., Blöcker,H. and Marky,L.A. (1986) Predicting DNA duplex stability from the base sequence. Proc. Natl. Acad. Sci. USA, 83, 3746-3750. SantaLucia,J.,Jr, Allawi,H.T. and Seneviratne,P.A. (1996) Improved nearest-neighbor parameters for predicting DNA duplex stability. Biochemistry, 35, 3555-3562. Sugimoto,N., Nakano,S., Yoneyama,M. and Honda,K. (1996) Improved thermodynamic parameters and helix initiation factor to predict stability of DNA duplexes. Nucleic Acids Res., 24, 4501-4505.