Preservatives from food-For food: Pea protein hydrolysate as a novel bio-preservative against Escherichia coli O157:H7 on a lettuce leaf.

Fresh-cut fruits and vegetables are becoming particularly popular as healthy fast-food options; however, they present challenges such as accelerated rates of decay and increased risk for contamination when compared to whole produce. Given that food safety must remain paramount for producers and manufacturers, research into novel, natural food preservation solutions which can help to ensure food safety and protect against spoilage is on the rise. In this work, we investigated the potential of using a novel protein hydrolysate, produced by the enzymatic hydrolysis of Pisum sativum (PSH), as a novel bio-preservative and its abilities to reduce populations of Escherichia coli O157:H7 after inoculation on a lettuce leaf. While unhydrolyzed P. sativum proteins show no antimicrobial activity, once digested, and purified, the enzymatically released peptides induced in vitro bactericidal effects on the foodborne pathogen at 8 mg/ml. When applied on an infected lettuce leaf, the PSH significantly reduced the number of bacteria recovered after 2 hr of treatment. PSH may be preferred over other preservation strategies based on its natural, inexpensive, sustainable source, environmentally friendly process, nontoxic nature, good batch to batch consistency, and ability to significantly reduce counts of E. coli both in vitro and in a lettuce leaf.

An Artificial-Intelligence-Discovered Functional Ingredient, NRT_N0G5IJ, Derived from Pisum sativum, Decreases HbA1c in a Prediabetic Population.

The prevalence of prediabetes is rapidly increasing, and this can lead to an increased risk for individuals to develop type 2 diabetes and associated diseases. Therefore, it is necessary to develop nutritional strategies to maintain healthy glucose levels and prevent glucose metabolism dysregulation in the general population. Functional ingredients offer great potential for the prevention of various health conditions, including blood glucose regulation, in a cost-effective manner. Using an artificial intelligence (AI) approach, a functional ingredient, NRT_N0G5IJ, was predicted and produced from Pisum sativum (pea) protein by hydrolysis and then validated. Treatment of human skeletal muscle cells with NRT_N0G5IJ significantly increased glucose uptake, indicating efficacy of this ingredient in vitro. When db/db diabetic mice were treated with NRT_N0G5IJ, we observed a significant reduction in glycated haemoglobin (HbA1c) levels and a concomitant benefit on fasting glucose. A pilot double-blinded, placebo controlled human trial in a population of healthy individuals with elevated HbA1c (5.6% to 6.4%) showed that HbA1c percentage was significantly reduced when NRT_N0G5IJ was supplemented in the diet over a 12-week period. Here, we provide evidence of an AI approach to discovery and demonstrate that a functional ingredient identified using this technology could be used as a supplement to maintain healthy glucose regulation.

Characterising the efficacy and bioavailability of bioactive peptides identified for attenuating muscle atrophy within a Vicia faba-derived functional ingredient.

Characterising key components within functional ingredients as well as assessing efficacy and bioavailability is an important step in validating nutritional interventions. Machine learning can assess large and complex data sets, such as proteomic data from plants sources, and so offers a prime opportunity to predict key bioactive components within a larger matrix. Using machine learning, we identified two potentially bioactive peptides within a Vicia faba derived hydrolysate, NPN_1, an ingredient which was previously identified for preventing muscle loss in a murine disuse model. We investigated the predicted efficacy of these peptides in vitro and observed that HLPSYSPSPQ and TIKIPAGT were capable of increasing protein synthesis and reducing TNF-α secretion, respectively. Following confirmation of efficacy, we assessed bioavailability and stability of these predicted peptides and found that as part of NPN_1, both HLPSYSPSPQ and TIKIPAGT survived upper gut digestion, were transported across the intestinal barrier and exhibited notable stability in human plasma. This work is a first step in utilising machine learning to untangle the complex nature of functional ingredients to predict active components, followed by subsequent assessment of their efficacy, bioavailability and human plasma stability in an effort to assist in the characterisation of nutritional interventions.

Discovery through Machine Learning and Preclinical Validation of Novel Anti-Diabetic Peptides.

While there have been significant advances in drug discovery for diabetes mellitus over the past couple of decades, there is an opportunity and need for improved therapies. While type 2 diabetic patients better manage their illness, many of the therapeutics in this area are peptide hormones with lengthy sequences and a molecular structure that makes them challenging and expensive to produce. Using machine learning, we present novel anti-diabetic peptides which are less than 16 amino acids in length, distinct from human signalling peptides. We validate the capacity of these peptides to stimulate glucose uptake and Glucose transporter type 4 (GLUT4) translocation in vitro. In obese insulin-resistant mice, predicted peptides significantly lower plasma glucose, reduce glycated haemoglobin and even improve hepatic steatosis when compared to treatments currently in use in a clinical setting. These unoptimised, linear peptides represent promising candidates for blood glucose regulation which require further evaluation. Further, this indicates that perhaps we have overlooked the class of natural short linear peptides, which usually come with an excellent safety profile, as therapeutic modalities.

An Artificial Intelligence Characterised Functional Ingredient, Derived from Rice, Inhibits TNF-α and Significantly Improves Physical Strength in an Inflammaging Population.

Food-derived bioactive peptides offer great potential for the treatment and maintenance of various health conditions, including chronic inflammation. Using in vitro testing in human macrophages, a rice derived functional ingredient natural peptide network (NPN) significantly reduced Tumour Necrosis Factor (TNF)-α secretion in response to lipopolysaccharides (LPS). Using artificial intelligence (AI) to characterize rice NPNs lead to the identification of seven potentially active peptides, the presence of which was confirmed by liquid chromatography tandem mass spectrometry (LC-MS/MS). Characterization of this network revealed the constituent peptides displayed anti-inflammatory properties as predicted in vitro. The rice NPN was then tested in an elderly "inflammaging" population with a view to subjectively assess symptoms of digestive discomfort through a questionnaire. While the primary subjective endpoint was not achieved, analysis of objectively measured physiological and physical secondary readouts showed clear significant benefits on the ability to carry out physical challenges such as a chair stand test that correlated with a decrease in blood circulating TNF-α. Importantly, the changes observed were without additional exercise or specific dietary alterations. Further health benefits were reported such as significant improvement in glucose control, a decrease in serum LDL concentration, and an increase in HDL concentration; however, this was compliance dependent. Here we provide in vitro and human efficacy data for a safe immunomodulatory functional ingredient characterized by AI.

Preclinical Evaluation of a Food-Derived Functional Ingredient to Address Skeletal Muscle Atrophy.

Skeletal muscle is the metabolic powerhouse of the body, however, dysregulation of the mechanisms involved in skeletal muscle mass maintenance can have devastating effects leading to many metabolic and physiological diseases. The lack of effective solutions makes finding a validated nutritional intervention an urgent unmet medical need. In vitro testing in murine skeletal muscle cells and human macrophages was carried out to determine the effect of a hydrolysate derived from vicia faba (PeptiStrong: NPN_1) against phosphorylated S6, atrophy gene expression, and tumour necrosis factor alpha (TNF-α) secretion, respectively. Finally, the efficacy of NPN_1 on attenuating muscle waste in vivo was assessed in an atrophy murine model. Treatment of NPN_1 significantly increased the phosphorylation of S6, downregulated muscle atrophy related genes, and reduced lipopolysaccharide-induced TNF-α release in vitro. In a disuse atrophy murine model, following 18 days of NPN_1 treatment, mice exhibited a significant attenuation of muscle loss in the soleus muscle and increased the integrated expression of Type I and Type IIa fibres. At the RNA level, a significant upregulation of protein synthesis-related genes was observed in the soleus muscle following NPN_1 treatment. In vitro and preclinical results suggest that NPN_1 is an effective bioactive ingredient with great potential to prolong muscle health.

Human milk and infant formulae: Peptide differences and the opportunity to address the functional gap.

Bovine-derived formula milk (FM) is a common substitute to human milk (HM), but lacks key functional benefits associated with HM. Accordingly, there have been significant efforts to humanise FM. Recent research has demonstrated that HM-derived peptides convey an array of beneficial bioactivities. Given that peptides serve as important signalling molecules offering high specificity and potency, they represent a prime opportunity to humanise FM. To further understand how HM-derived peptides contribute to infant health, we used peptidomics and bioinformatics to compare the peptide profile of HM to commercially available FM. We found clear and substantial differences between HM and FM in terms of peptide physicochemical properties, protein coverage and abundance. We additionally identified 618 peptides specific to HM that represent an important untapped source to be explored for novel bioactivities. While further study is required, our findings highlight the potential of a peptide-based approach to address the functional gap in FM.

Reconstructing protein structures by neural network pairwise interaction fields and iterative decoy set construction.

Predicting the fold of a protein from its amino acid sequence is one of the grand problems in computational biology. While there has been progress towards a solution, especially when a protein can be modelled based on one or more known structures (templates), in the absence of templates, even the best predictions are generally much less reliable. In this paper, we present an approach for predicting the three-dimensional structure of a protein from the sequence alone, when templates of known structure are not available. This approach relies on a simple reconstruction procedure guided by a novel knowledge-based evaluation function implemented as a class of artificial neural networks that we have designed: Neural Network Pairwise Interaction Fields (NNPIF). This evaluation function takes into account the contextual information for each residue and is trained to identify native-like conformations from non-native-like ones by using large sets of decoys as a training set. The training set is generated and then iteratively expanded during successive folding simulations. As NNPIF are fast at evaluating conformations, thousands of models can be processed in a short amount of time, and clustering techniques can be adopted for model selection. Although the results we present here are very preliminary, we consider them to be promising, with predictions being generated at state-of-the-art levels in some of the cases.

SCLpredT: Ab initio and homology-based prediction of subcellular localization by N-to-1 neural networks.

The prediction of protein subcellular localization is a important step towards the prediction of protein function, and considerable effort has gone over the last decade into the development of computational predictors of protein localization. In this article we design a new predictor of protein subcellular localization, based on a Machine Learning model (N-to-1 Neural Networks) which we have recently developed. This system, in three versions specialised, respectively, on Plants, Fungi and Animals, has a rich output which incorporates the class "organelle" alongside cytoplasm, nucleus, mitochondria and extracellular, and, additionally, chloroplast in the case of Plants. We investigate the information gain of introducing additional inputs, including predicted secondary structure, and localization information from homologous sequences. To accommodate the latter we design a new algorithm which we present here for the first time. While we do not observe any improvement when including predicted secondary structure, we measure significant overall gains when adding homology information. The final predictor including homology information correctly predicts 74%, 79% and 60% of all proteins in the case of Fungi, Animals and Plants, respectively, and outperforms our previous, state-of-the-art predictor SCLpred, and the popular predictor BaCelLo. We also observe that the contribution of homology information becomes dominant over sequence information for sequence identity values exceeding 50% for Animals and Fungi, and 60% for Plants, confirming that subcellular localization is less conserved than structure. SCLpredT is publicly available at http://distillf.ucd.ie/sclpredt/. Sequence- or template-based predictions can be obtained, and up to 32kbytes of input can be processed in a single submission.

Accurate prediction of protein enzymatic class by N-to-1 Neural Networks.

We present a novel ab initio predictor of protein enzymatic class. The predictor can classify proteins, solely based on their sequences, into one of six classes extracted from the enzyme commission (EC) classification scheme and is trained on a large, curated database of over 6,000 non-redundant proteins which we have assembled in this work. The predictor is powered by an ensemble of N-to-1 Neural Network, a novel architecture which we have recently developed. N-to-1 Neural Networks operate on the full sequence and not on predefined features. All motifs of a predefined length (31 residues in this work) are considered and are compressed by an N-to-1 Neural Network into a feature vector which is automatically determined during training. We test our predictor in 10-fold cross-validation and obtain state of the art results, with a 96% correct classification and 86% generalized correlation. All six classes are predicted with a specificity of at least 80% and false positive rates never exceeding 7%. We are currently investigating enhanced input encoding schemes which include structural information, and are analyzing trained networks to mine motifs that are most informative for the prediction, hence, likely, functionally relevant.