Here, cluster analysis showed that a database of 158 peptides formed 21 clusters based on net positive charge, hydrophobicity and amphiphilicity. In general these clusters showed similar median toxicities (p = 0.176) against eukaryotic cell lines and no single combination of these properties was found optimal for efficacy. The database contained 14 peptides, which showed selectivity for tumour cell lines only (ACP CT ), 123 peptides with general toxicity to eukaryotic cells (ACP GT ) and 21 inactive peptides (ACP I ). Hydrophobic arc size analysis showed that there was no significant difference across the datasets. Even though there was no correlation there was no correlation observed, peptides with wide hydrophobic arcs (> 270°) appeared less toxic. Extended hydrophobic moment plot analysis predicted that over 50% of ACP CT and ACP GT peptides would be surface active, which led to the suggestion that amphiphilicity is a key driver of the membrane interactions for these peptides but probably plays a role in their efficacy rather than their selectivity. This analysis also predicted that only 14% of ACP CT peptides compared to 45% of ACP GT peptides were candidates for tilted peptide formation. This implies that those peptides with nonspecific activity may have a tendency towards the utilisation of membrane disruptive structures such as tilt peptides which led to the suggestion that the absence of this structure may support cancer cell selectivity. However, these analyses predicted that ACP I peptides, which possess no anticancer activity, would also form surface active and tilted α-helices, clearly showing that other factors are involved in determining the efficacy and selectivity of ACPs.