Anthracycline antibiotics such as doxorubicin and its analogues have been in common use as anticancer drugs for almost half a century. There has been intense interest in the DNA binding sequence specificity of these compounds in recent years, with the hope that a compound could be identified that could possibly modulate gene expression or exhibit reduced toxicity. To computationally analyze this phenomenon, we have constructed molecular models of 65 doxorubicin analogues and their complexes with eight distinct DNA octamer sequences. The HINT (Hydropathic INTeractions) program was utilized to describe binding, including differences in the functional group contributions as well as sequence selectivity. Of these 65 compounds, two compounds were calculated to have a selectivity (the calculated DeltaDeltaG(sel) between the sequence with the strongest binding and the second strongest binding sequence) greater than -0.75 kcal mol(-1) for one sequence over all others, 10 compounds were specific between -0.50 and -0.74 kcal mol(-1), 18 compounds were specific between -0.25 and -0.49 kcal mol(-1), and 35 compounds were virtually nonspecific with a DeltaDeltaG below -0.24 kcal mol(-1). Several compounds have been identified from this study that include features which may enhance sequence selectivity, including several with a halogen in lieu of the 4'-OH in the daunosamine sugar, one compound with a nonaromatic six-membered ring (pirarubicin) in place of the 4'-OH, and a compound with an aromatic ring in the vicinity of the C(14) region (zorubicin). Removal of the methoxy group at the C(4) position on the aglycone portion also appears to add potency and selectivity (idarubicin). Overall, efficient computational methods are presented that can be utilized to analyze the free energy of binding and sequence selectivity of both known and designed analogues of doxorubicin to identify future lead compounds for further experimental research.