Accuracy benchmark table All sequences used in the accuracy benchmark, along with the experimental and calculated data can be downloaded here. The source of the experimental data available for these sequences is shown in the table. For an explanation of the consensus type field ('CON_TYPE') see the Figure and read the following section. For more details read the paper: Panjkovich, A. and Melo, F. (2004). Bioinformatics (in press). Distribution of the available experimental data The total amounts of oligonucleotides with available experimental Tm values (listed in the table) have been mapped into this grid according to CG-content and length. The consensus among two or three methods is defined when at least 80 percent of the sequences exhibit an absolute difference between the calculated Tm values of less than 5° C. All pairwise comparisons were carried out, as well as the simultaneous comparison of the three thermodynamic methods. Th1 (Breslauer et al., 1986) and Th2 (SantaLucia et al., 1996) did not show similar behaviour in the whole range of sequence length and percentage of CG-content. The observed consensus among the methods is as follows: Simultaneously, Th1 and Th3, and Th2 and Th3, exhibit similar values (white color); only Th1 (Breslauer et al., 1986) and Th3 (Sugimoto et al., 1996) exhibit similar values (light gray); only Th2 (SantaLucia et al., 1996) and Th3 (Sugimoto et al., 1996) exhibit similar values (dark gray); and finally, no consensus is observed among any of the methods (black).