The complex pathophysiology of human immunodeficiency virus (HIV) infection and the relatively high mutation rate of the retrovirus make it challenging to design effective anti-HIV vaccines. Several attempts have been made during the last decades to elucidate the enigmatic immunology of HIV infection and to predict potential immunogenic peptides for active vaccination using bioinformatic analysis methods. The results obtained to date to address this important problem are scarce. In this study, we exploit available HIV databases and analyse previously characterized HIV-encoded linear B-cell epitopes for their amino acid sequence similarity to the human or murine host proteome. We obtained further documentation that the HIV-derived antibody-targeted sequences mostly coincide with peptide areas rarely shared with the host proteins. In toto, our past and present data give clear-cut support to the statement that low-similarity to the host proteome is a major mechanism in defining viral peptide immunogenicity and indicate a possible way for inducing effective, high-titer, and non-crossreactive antibodies to be used in anti-HIV vaccine therapy.