Machine learning (ML) consists of the recognition of patterns from training data and offers the opportunity to exploit large structure-activity databases for drug design. In the area of peptide drugs,...

<p>Machine learning (ML) consists in the recognition of patterns from training data and offers the opportunity to exploit large structure-activity database sets for drug design. In the area of peptide drugs, ML is mostly being tested to design antimicrobial peptides (AMPs), a class of biomolecules potentially useful to fight multidrug resistant bacteria. ML models have successfully identified membrane disruptive amphiphilic AMPs, however without addressing the associated toxicity to human red blood cells. Here we trained recurrent neural networks (RNN) with data from DBAASP (Database of Antimicrobial Activity and Structure of Peptides) to design short non-hemolytic AMPs. Synthesis and testing of 28 generated peptides, each at least 5 mutations away from training data, allowed us to identify eight new non-hemolytic AMPs against <i>Pseudomonas aeruginosa</i>, <i>Acinetobacter baumannii</i>, and methicillin resistant<i> Staphylococcus aureus</i> (MRSA). These results show that machine learning (ML) can be used to design new non-hemolytic AMPs.</p>

The peptide α-helix is right-handed when containing amino acids with L-chirality, and left-handed with D-chirality, however mixed chirality peptides generally do not form α-helices unless the non-natural residue amino-isobutyric acid...

Herein, we report dipropylamine (DPA) as a fluorenylmethyloxycarbonyl (Fmoc) deprotection reagent to strongly reduce aspartimide formation compared to piperidine (PPR) in high-temperature (60 °C) solid-phase peptide synthesis (SPPS). In contrast to PPR, DPA is readily available, inexpensive, low toxicity, and nonstench. DPA also provides good yields in SPPS of non-aspartimide-prone peptides and peptide dendrimers.

<p>Machine learning (ML) consists in the recognition of patterns from training data and offers the opportunity to exploit large structure-activity database sets for drug design. In the area of peptide drugs, ML is mostly being tested to design antimicrobial peptides (AMPs), a class of biomolecules potentially useful to fight multidrug resistant bacteria. ML models have successfully identified membrane disruptive amphiphilic AMPs, however without addressing the associated toxicity to human red blood cells. Here we trained recurrent neural networks (RNN) with data from DBAASP (Database of Antimicrobial Activity and Structure of Peptides) to design short non-hemolytic AMPs. Synthesis and testing of 28 generated peptides, each at least 5 mutations away from training data, allowed us to identify eight new non-hemolytic AMPs against <i>Pseudomonas aeruginosa</i>, <i>Acinetobacter baumannii</i>, and methicillin resistant<i> Staphylococcus aureus</i> (MRSA). These results show that machine learning (ML) can be used to design new non-hemolytic AMPs.</p>