In a recent paper, we have described the discovery of antimalarial compounds derived from tadalafil, using a drug-to-genome-to-drug approach ( J. Med. Chem. 2011 , 54 ( 9 ), pp 3222 - 3240 ). We have shown that these derivatives inhibit the phosphodiesterase activity of Plasmodium falciparum and the parasite growth in culture. In this paper, we describe the optimization of these compounds. A direct consequence of our approach based on gene orthology is the lack of selectivity of the compounds over the original activity on the human target. We demonstrate here that it is possible to take advantage of subtle differences in SAR between HsPDE5 inhibition and antiplasmodial activity to improve significantly the selectivity. In particular, the replacement of the piperonyl group in compound 2 by a dimethoxyphenyl group was the best way to optimize selectivity. This observation is consistent with the differences between human and plasmodial sequences in the Q2 pocket receiving this group.