Quantitative sequence–activity modeling of antimicrobial hexapeptides using a segmented principal component strategy: an approach to describe and predict activities of peptide drugs containing l/d and unnatural residues

On QSPR analysis of glaucoma drugs using machine learning with XGBoost and regression models

Glaucoma is an irreversible, progressive, degenerative eye disorder arising because of increased intraocular pressure, resulting in eventual vision loss if untreated. The QSPR relates, mathematically, by employing various algorithms, a specified property of a molecule that arises either from physical, chemical, or biological phenomena using various aspects of its structure. Here in, a similar application based on topological indices and inferences derived from the structure for the calculation of different drug properties like molar refractivity, refractive index, enthalpy, boiling points, molecular weight, and polarizability is presented. Linear regression is developed between the features of QSPR, coupled with topological indices, and performance assessment is conducted in conjunction with an Extreme Gradient Boosting model. From the obtained results, one can draw out that the XGBoost model will give better results than those from simple regression methods, specifically for polarizability. Predicted values validate the experiments. This work shows how the combination of machine learning techniques and topological indices can be used to get the best out of predictive modeling.

Unnatural amino acids: promising implications for the development of new antimicrobial peptides

The increasing incidence and rapid spread of bacterial resistance to conventional antibiotics are a serious global threat to public health, highlighting the need to develop new antimicrobial alternatives. Antimicrobial peptides (AMPs) represent a class of promising natural antibiotic candidates due to their broad-spectrum activity and low tendency to induce resistance. However, the development of AMPs for medical use is hampered by several obstacles, such as moderate activity, lability to proteolytic degradation, and low bioavailability. To date, many researchers have focussed on the optimization or design of novel artificial AMPs with desired properties. Unnatural amino acids (UAAs) are valuable building blocks in the manufacture of a variety of pharmaceuticals, and have been used to develop artificial AMPs with specific structural and physicochemical properties. Rational incorporation of UAAs has become a very promising approach to endow AMPs with strong and long-lasting activity but no toxicity. This review aims to summarize key approaches that have been used to incorporate UAAs to develop novel AMPs with improved properties and better performance. It is anticipated that this review will guide future design considerations for UAA-based antimicrobial applications.

From biomedicinal to in silico models and back to therapeutics: a review on the advancement of peptidic modeling

Bioactive peptides participate in numerous metabolic functions of living organisms and have emerged as potential therapeutics on a diverse range of diseases. Albeit peptide design does not go without challenges, overwhelming advancements on in silico methodologies have increased the scope of peptide-based drug design and discovery to an unprecedented amount. Within an in silico model versus an experimental validation scenario, this review aims to summarize and discuss how different in silico techniques contribute at present to the design of peptide-based molecules. Published in silico results from 2014 to 2018 were selected and discriminated in major methodological groups, allowing a transversal analysis, promoting a landscape vision and asserting its increasing value in drug design.

Quantitative sequence-activity modeling of ACE peptide originated from milk using ACC-QTMS amino acid indices

Up to now, numerous peptides/hydrolysates derived from casein and whey protein have shown angiotensin-I-converting enzyme (ACE) inhibitory. In this research, quantum topological molecular similarity (QTMS) indices of amino acids were utilized in quantitative sequence-activity modeling (QSAM) to predict the activity of a set of milk-driven peptides with ACE inhibition. Since the derived peptides have not the same number of residues, we overcame this issue by auto cross covariance (ACC) methodology. Then, some QSAMs were built to predict the pIC₅₀ value of ACE peptides derived from Bovine Casein and Whey. The model established an acceptable relationship between the selected variables and the pIC₅₀ of the peptides. To estimate the performance of the developed models, casein and whey proteins from human, goat, bovine and sheep were virtually broken by trypsin and chymotrypsin enzymes and the ACE activity of the resultant virtual peptides were predicted and some new ACE peptides were proposed.