Multifunctional Peptide Research Articles

Antimicrobial Peptides (AMPs) and the Microbiome in Preterm Infants: Consequences and Opportunities for Future Therapeutics.

Antimicrobial peptides (AMPs) are crucial components of the innate immune system in various organisms, including humans. Beyond their direct antimicrobial effects, AMPs play essential roles in various physiological processes. They induce angiogenesis, promote wound healing, modulate immune responses, and serve as chemoattractants for immune cells. AMPs regulate the microbiome and combat microbial infections on the skin, lungs, and gastrointestinal tract. Produced in response to microbial signals, AMPs help maintain a balanced microbial community and provide a first line of defense against infection. In preterm infants, alterations in microbiome composition have been linked to various health outcomes, including sepsis, necrotizing enterocolitis, atopic dermatitis, and respiratory infections. Dysbiosis, or an imbalance in the microbiome, can alter AMP profiles and potentially lead to inflammation-mediated diseases such as chronic lung disease and obesity. In the following review, we summarize what is known about the vital role of AMPs as multifunctional peptides in protecting newborn infants against infections and modulating the microbiome and immune response. Understanding their roles in preterm infants and high-risk populations offers the potential for innovative approaches to disease prevention and treatment.

Ultrasensitive electrochemical sensor based on synergistic effect of Ag@MXene and antifouling cyclic multifunctional peptide for PD-L1 detection in serum.

Asensing interface co-constructed from the two-dimensional conductive material (Ag@MXene) and an antifouling cyclic multifunctional peptide (CP)is described. While the large surface area of Ag@MXene loads more CP probes, CP binds to Ag@MXene to form a fouling barrier and ensure the structural rigidity of the targeting sequence. This strategy synergistically enhances the biosensor's sensitivity and resistance to contamination. The SPR results showed that the binding affinity of the CP to the target was 6.23 times higher than that of the antifouling straight-chain multifunctional peptide (SP) to the target. In the 10mg/mL BSA electrochemical fouling test, the fouling resistance of Ag@MXene + CP (composite sensing interface of CP combined with Ag@MXene) was 30 times higher than that of the bare electrode. The designed electrochemical sensor exhibited good selectivity and wide dynamic response range at PD-L1 concentrations from 0.1 to 50 ng/mL. The lowest detection limit was 24.54 pg/mL (S/N = 3). Antifouling 2D materials with a substantial specific surface area, coupled with non-straight chain antifouling multifunctional peptides, offer a wide scope for investigating the sensitivity and antifouling properties of electrochemical sensors.

Introducing Bacillus natto and Propionibacterium shermanii into soymilk fermentation: A promising strategy for quality improvement and bioactive peptide production during in vitro digestion

Herein, the texture properties, polyphenol contents, and in vitro protein digestion characteristics of soymilk single- or co-fermented by non-typical milk fermenter Bacillus natto (B. natto), Propionibacterium freudenreichii subsp. shermanii (P. shermanii), and traditional milk fermenter were evaluated. Co-fermenting procedure containing B. natto or P. shermanii could raise the amounts of gallic acid, caffeic acid, and GABA when compared to the unfermented soymilk. Co-fermented soymilk has higher in vitro protein digestibility and nutritional protein quality. Through peptidomic analysis, the co-work of P. shermanii and Lactobacillus plantarum (L. plantarum) may release the highest relative percentage of bioactive peptides, while the intervention of B. natto and Streptococcus thermophilus (S. thermophilus) resulted in more differentiated peptides. The multi-functional bioactive peptides were mainly released from glycine-rich protein, β-conglycinin alpha subunit 1, and ACB domain-containing protein. These findings indicated the potential usage of B. natto/S. thermophilus or P. shermanii/L. plantarum in bio-enhanced soymilk fermentation.

CELA-MFP: a contrast-enhanced and label-adaptive framework for multi-functional therapeutic peptides prediction.

Functional peptides play crucial roles in various biological processes and hold significant potential in many fields such as drug discovery and biotechnology. Accurately predicting the functions of peptides is essential for understanding their diverse effects and designing peptide-based therapeutics. Here, we propose CELA-MFP, a deep learning framework that incorporates feature Contrastive Enhancement and Label Adaptation for predicting Multi-Functional therapeutic Peptides. CELA-MFP utilizes a protein language model (pLM) to extract features from peptide sequences, which are then fed into a Transformer decoder for function prediction, effectively modeling correlations between different functions. To enhance the representation of each peptide sequence, contrastive learning is employed during training. Experimental results demonstrate that CELA-MFP outperforms state-of-the-art methods on most evaluation metrics for two widely used datasets, MFBP and MFTP. The interpretability of CELA-MFP is demonstrated by visualizing attention patterns in pLM and Transformer decoder. Finally, a user-friendly online server for predicting multi-functional peptides is established as the implementation of the proposed CELA-MFP and can be freely accessed at http://dreamai.cmii.online/CELA-MFP.

A Super-Antifouling Electrochemical Biosensor for Protein Detection in Complex Biofluids Based on PEGylated Multifunctional Peptide.

Overcoming the influence of interfering substances in the environment and achieving superior sensing performance are significant challenges in biomarker detection within complex matrices. Herein, an integrated electrochemical sensing platform for sensitive detection of biomarkers in complex biofluids was developed based on a newly designed PEGylated multifunctional peptide (PEG-MPEP). The designed PEG-MPEP contains a poly(serine) sequence (-ssssss-) as the antifouling part and recognition peptide sequence (-avwgrwh) specific for the target human immunoglobulin G (IgG). To improve the peptide stability to protease hydrolysis, d-amino acids were adopted to synthesize the whole peptide. Additionally, the PEGylation can further enhance the stability of the peptide, and the PEG itself was also antifouling, ensuring superstrong antifouling capability of the PEG-MPEP. The designed PEG-MPEP-based biosensor possessed a high sensitivity for the detection of IgG in the range of 1.0 pg mL-1 to 1.0 μg mL-1, with a low limit of detection (0.41 pg mL-1), and it was capable of assaying targets accurately in real serum samples. Compared with conventional peptide-modified biosensors, the PEG-MPEP-modified biosensor exhibited superior antifouling and antihydrolysis properties in complex biofluid, showcasing promising potential for practical assay applications.

Adrenomedullin improved endothelial dysfunction via receptor-Akt pathway in rats with obesity-related hypertension.

Obesity-related hypertension (OH) is accompanied by obvious endothelial dysfunction, which contributes to increased peripheral vascular resistance and hypertension. Adrenomedullin (ADM), a multifunctional active peptide, is elevated in obese humans. The OH rats induced by high fat diet (HFD) for 28 weeks and the human umbilical vein endothelial cells (HUVECs)-treated by palmitic acid (PA) were used to investigate the effects of ADM on endothelial dysfunction and the underlying mechanisms. Vascular reactivity was assessed using mesenteric arteriole rings, and the protein expression levels were examined by Western blot analysis. Compared with the control rats, OH rats exhibited hypertension and endothelial dysfunction, along with reduced eNOS protein expression and Akt activation, and increased protein expression of proinflammatory cytokines and ROS levels. Four-week ADM administration improved hypertension and endothelial function, increased eNOS protein expression and Akt activation, and attenuated endothelial inflammation and oxidative stress in OH rats. In vitro experiment, the antagonism of ADM receptors with ADM22-52 and the suppression of Akt signaling with A6730 significantly blocked ADM-caused increase of NO content and activation of eNOS and Akt, and inhibited the anti-inflammatory and anti-oxidant effect of ADM in PA-stimulated HUVECs. These data indicate that endothelial dysfunction in OH rats is partially attributable to the decreased NO level, and the increased inflammation and oxidative stress. ADM improves endothelial function and exerts hypotensive effect depending on the increase of NO, and its anti-inflammatory and anti-oxidant effect via receptor-Akt pathway.

An Effective Modification Strategy to Build Multifunctional Peptides Based on a Trypsin Inhibitory Peptide of the Kunitz Family.

Peptides with antimicrobial activity or protease inhibitory activity are potential candidates to supplement traditional antibiotics or cancer chemotherapies. However, the potential of many peptides are limited by drawbacks such as cytotoxicity or susceptibility to hydrolysis. Therefore, strategies to modify the structure of promising peptides may represent an effective approach for developing more promising clinical candidates. In this study, the mature peptide OSTI-1949, a Kunitz-type inhibitor from Odorrana schmackeri, and four designed analogues were successfully synthesised. In contrast to the parent peptide, the analogues showed impressive multi-functionality including antimicrobial, anticancer, and trypsin inhibitory activities. In terms of safety, there were no obvious changes observed in the haemolytic activity at the highest tested concentration, and the analogue OSTI-2461 showed an increase in activity against cancer cell lines without cytotoxicity to normal cells (HaCaT). In summary, through structural modification of a natural Kunitz-type peptide, the biological activity of analogues was improved whilst retaining low cytotoxicity. The strategy of helicity enhancement by forming an artificial α-helix and ß-sheet structure provides a promising way to develop original bioactive peptides for clinical therapeutics.

Multifunctional host-defense peptides isolated from skin secretions of the banana tree dwelling frog Boana platanera (Hylidae; Hylinae)

Five host-defense peptides (figainin 2PL, hylin PL, raniseptin PL, plasticin PL, and peptide YL) were isolated from norepinephrine-stimulated skin secretions of the banana tree dwelling frog Boana platanera (Hylidae; Hylinae) collected in Trinidad.Raniseptin PL (GVFDTVKKIGKAVGKFALGVAKNYLNS.NH2) and figainin 2PL (FLGTVLKLGKAIAKTVVPMLTNAMQPKQ. NH2) showed potent and rapid bactericidal activity against a range of clinically relevant Gram-positive and Gram-negative ESKAPE + pathogens and Clostridioides difficile. The peptides also showed potent cytotoxic activity (LC50 values < 30 μM) against A549, MDA-MB-231 and HT29 human tumor-derived cell lines but appreciably lower hemolytic activity against mouse erythrocytes (LC50 = 262 ± 14 μM for raniseptin PL and 157 ± 16 μM for figainin 2PL). Hylin PL (FLGLIPALAGAIGNLIK.NH2) showed relatively weak activity against microorganisms but was more hemolytic. The glycine-leucine-rich peptide with structural similarity to the plasticins (GLLSTVGGLVGGLLNNLGL.NH2) and the non-cytotoxic peptide YL (YVPGVIESLL.NH2) lacked antimicrobial and cytotoxic activities. Hylin PL, raniseptinPL and peptide YL stimulated the rate of release of insulin from BRIN-BD11 clonal β-cells at concentrations ≥100 nM. Peptide YL was the most effective (2.3-fold increase compared with basal rate at 1 μM concentration) and may represent a template for the design of a new class of incretin-based anti-diabetic drugs.

A non-viral DNA delivery system consisting of multifunctional chimeric peptide fused with zinc-finger protein

Non-viral gene delivery systems have received sustained attention as a promising alternative to viral vectors for disease treatment and prevention in recent years. Numerous methods have been developed to enhance gene uptake and delivery in the cytoplasm; however, due to technical difficulties and delivery efficiency,these systems still face challenges in a range of biological applications, especially invivo. To alleviate this challenge, we devised a novel system for gene delivery based on a recombinant protein eTAT-ZF9-NLS, which consisted of a multifunctional chimeric peptide and a zinc-finger protein with sequence-specific DNA-binding activity. High transfection efficiency was observed in several mammalian cells after intracellular delivery of plasmid containing ZF9-binding sites mediated by eTAT-ZF9-NLS. Our new approach provides a novel transfection strategy and the transfection efficiency was confirmed both invitro and invivo, making it a preferential transfection reagent for possible gene therapy.

MMDB: Multimodal dual-branch model for multi-functional bioactive peptide prediction

Bioactive peptides can hinder oxidative processes and microbial spoilage in foodstuffs and play important roles in treating diverse diseases and disorders. While most of the methods focus on single-functional bioactive peptides and have obtained promising prediction performance, it is still a significant challenge to accurately detect complex and diverse functions simultaneously with the quick increase of multi-functional bioactive peptides. In contrast to previous research on multi-functional bioactive peptide prediction based solely on sequence, we propose a novel multimodal dual-branch (MMDB) lightweight deep learning model that designs two different branches to effectively capture the complementary information of peptide sequence and structural properties. Specifically, a multi-scale dilated convolution with Bi-LSTM branch is presented to effectively model the different scales sequence properties of peptides while a multi-layer convolution branch is proposed to capture structural information. To the best of our knowledge, this is the first effective extraction of peptide sequence features using multi-scale dilated convolution without parameter increase. Multimodal features from both branches are integrated via a fully connected layer for multi-label classification. Compared to state-of-the-art methods, our MMDB model exhibits competitive results across metrics, with a 9.1% Coverage increase and 5.3% and 3.5% improvements in Precision and Accuracy, respectively.

The State-of-the-Art Mechanisms and Antitumor Effects of Somatostatin in Colorectal Cancer: A Review.

Somatostatin, a somatotropin release inhibiting factor (SST, SRIF), is a widely distributed multifunctional cyclic peptide and acts through a transmembrane G protein-coupled receptor (SST1-SST5). Over the past decades, research has begun to reveal the molecular mechanisms underlying the anticancer activity of this hormonal peptide. Among gastrointestinal tract (GIT) tumors, direct and indirect antitumor effects of SST have been documented best in gastroenteropancreatic neuroendocrine tumors (GEP-NETs) and less well in non-endocrine cancers, including sporadic colorectal cancer (CRC). In the latter, the signaling pathways involved in the antitumor function of SST are primarily MAPK/ERK/AKT and Wnt/β-catenin. Direct (involving the MAPK pathway) and indirect (VEGF production) antiangiogenic effects of SST in CRC have also been described. The anti-inflammatory role of SST in CRC is emphasized, but detailed molecular mechanisms are still being explored. The role of SST in tumor genome/tumor microenvironment (TME)/host's gut microbiome interactions is only partially known. The results of SST analogues (SSAs)' treatment of sporadic CRC in monotherapy in vivo are not spectacular. The current review aims to present the state-of-the-art mechanisms and antitumor activity of endogenous SST and its synthetic analogues in CRC, with particular emphasis on sporadic CRC.

In silico bioactivity analysis of peptide fractions derived from brewer's spent grain hydrolysates

SummaryBrewer's spent grain is an agro‐industrial waste of barley malt from the manufacture of beer, which is rich in protein. Thus, this by‐product of the beer industry could be utilised to obtain bioactive peptides through enzymatic hydrolysis. The present study aimed to evaluate potential bioactivities (anti‐inflammatory, antidiabetic, antioxidant, antimicrobial) of peptide sequences identified in 5‐kDa peptide fractions derived from brewer's spent grain hydrolysates. Two experimental hydrolysates (run 5 and run 21) were selected based on their higher degree of hydrolysis (&gt;55%), which were subjected to peptide sequence identification and used for in silico analysis. A total of fifty peptide sequences were identified between both runs. In silico analysis identified seventeen and thirteen multifunctional peptides in runs 5 and 21, respectively, which showed between two and four potential bioactivities. Interestingly, one peptide sequence (DLPPER) showed four potential bioactivities and, according to the Docking analysis, showed better Vina scores towards the protein's targets DPP‐IV (antidiabetic) and glucocorticoid receptor (anti‐inflammatory). These results demonstrate that peptides from brewer's spent grain hydrolysates may contribute to improving human health by offering potential protection against some chronic‐related diseases; however, further follow‐up in vitro and in vivo studies are needed to evaluate its actual capability.

SME-MFP: A novel spatiotemporal neural network with multiangle initialization embedding toward multifunctional peptides prediction

As a promising alternative to conventional antibiotic drugs in the biomedical field, functional peptide has been widely used in disease treatment owing to its low toxicity, high absorption rate, and biological activity. Recently, several machine learning methods have been developed for functional peptide prediction. However, the main research heavily relies on statistical features and few consider multifunctional peptide identification. So, we propose SME-MFP, a novel predictor in the imbalanced multi-label functional peptide datasets. First, we employ physicochemical and evolutionary information to represent the peptide sequence’s initialization features from multiple perspectives. Second, the features are fused and then put into spatial feature extractors, where the residual connection and multiscale convolutional neural network extract more discriminative features of different lengths’ peptide sequences. Besides, we also design AFT-based temporal feature extractors to fully capture the global interactions of the sequences. Finally, devising a new loss to replace the traditional cross entropy loss to settle the class imbalance problems. The results show that our framework not only enhances the model's ability to capture sequence features effectively, but also accuracy improves by 3.89% over existing methods on public peptide datasets.

Tet1 peptide and zinc (II)-adenine multifunctional module functionalized polycations as efficient siRNA carriers for Parkinson's disease

A bioreducible Zn (II)-adenine multifunctional module (BS) and Tet1 peptide were used to modify low-molecular-weight PEI3.5k (polyethyleneimine with molecular weight of 3.5 kDa)into a siRNA vector Zn-PB-T with high transfection efficiency in neurons. A GSH-responsive breakable disulfide spacer was introduced into BS to realize the controlled release of siRNA from the polyplexes in cytoplasm. Zn-PB showed >90% transfection rates in multiple cell lines (3 T3, HK-2, HepG2, 293 T, HeLa, PANC-1)，and 1.8-folds higher EGFP knockdown rates than commercial Lipo2k in normal cell line 293 T and cancer cell line HepG2. And Zn-PB-T1 showed 4.7–4.9- and 8.0–8.1-folds higher transfection efficiency comparing to commercial Lipo2k and PEI25k (polyethyleneimine with molecular weight of 25 kDa) in PC12 cells respectively, 2.1-fold EGFP gene silencing efficiency (96.6% EGFP knockdown rates) superior to commercial Lipo2k in neurons. In Parkinson's model, Zn-PB-T1/SNCA-siRNA can effectively protect neurons against MPP+-induced cell death and apoptosis, increasing the cell survival rate to 84.6% and reducing the cell apoptosis rate to 10.8%. This work demonstrated the promising application prospects of the resulting efficient siRNA carriers in siRNA-mediated gene therapy of Parkinson's disease.

Multifunctional self-healing peptide hydrogel for wound healing

The complete healing of wounds remains a challenge in clinical care. In addition, various complications such as inflammation and infection that may occur during skin wound healing can impede the healing process. Here, we constructed a multifunctional self-repairing hydrogel by utilizing Schiff base bonds. This hydrogel exhibited good self-healing properties and could cope with destructive external influences. The self-healing hydrogel was injectable, ensuring that the hydrogel dressing adhered to the wound. Carboxymethyl chitosan and oxidized chondroitin sulfate demonstrated good biocompatibility and multiple bioactivities and were successfully used to prepare self-healing hydrogels. Meanwhile, the SIKVAV biopeptide was less expensive and more morphologically stable than vascular endothelial growth factor and had a high pro-angiogenic activity. Thus, the SIKVAV biopeptide was cross-linked to the oxidized chondroitin sulfate of the hydrogel through covalent bonding to avoid rapid biopeptide degradation, achieving a slow release of the drug. This peptide hydrogel exhibited good biocompatibility and antimicrobial properties; moreover, experiments conducted on mice revealed that it could effectively promote angiogenesis and skin tissue repair. These findings suggest that the injectable self-repairing peptide hydrogel may facilitate skin wound healing and other applications.

LL-37_Renalexin hybrid peptide exhibits antimicrobial activity at lower MICs than its counterpart single peptides

An alarming global public health and economic peril has been the emergence of antibiotic resistance resulting from clinically relevant bacteria pathogens, including Enterococcus faecium, Staphylococcus aureus, Klebsiella pneumonia, Acinetobacter baumannii, Pseudomonas aeruginosa, and Enterobacter species constantly exhibiting intrinsic and extrinsic resistance mechanisms against last-resort antibiotics like gentamycin, ciprofloxacin, tetracycline, colistin, and standard ampicillin prescription in clinical practices. The discovery and applications of antimicrobial peptides (AMPs) with antibacterial properties have been considered and proven as alternative antimicrobial agents to antibiotics. In this study, we have designed, produced, and purified a recombinant novel multifunctional hybrid antimicrobial peptide LL-37_Renalexin for the first time via the application of newly designed flexible GS peptide linker coupled with the use of our previously characterized small metal-binding proteins SmbP and CusF3H+ as carrier proteins that allow for an enhanced bacterial expression, using BL21(DE3) and SHuffle T7(DE3) Escherichia coli strains, and purification of the hybrid peptide via immobilized metal affinity chromatography. The purified tag-free LL-37_Renalexin hybrid peptide exhibited above 85% reduction in bacteria colony-forming units and broad-spectrum antimicrobial effects against Staphylococcus aureus, Escherichia coli, Methicillin-resistant Staphylococcus aureus (MRSA), and Klebsiella pneumoniae bacteria clinical isolates at a lower minimum inhibition concentration level (10–33 μM) as compared to its counterpart single-AMPs LL-37 and Renalexin (50–100 μM).Key points• The hybrid antimicrobial peptide LL-37_Renalexin has been designed using a GS linker.• The peptide was expressed with the carrier proteins SmbP and CusF3H+.• The hybrid peptide shows antibacterial potency against clinical bacterial isolates.

Construction of Bacillus subtilis for efficient production of fengycin from xylose through CRISPR-Cas9.

Fengycin is a multifunctional peptide antibiotic produced mainly by Bacillus species and the purpose of this research was to construct a Bacillus subtilis strain that can produce fengycin with the xylose as the substrate with CRSIPR-Cas9. Hence, at the beginning of this study, functional sfp and degQ were expressed in B. subtilis 168 strain to give the strain the ability to produce the fengycin with the titer of 71.21 mg/L. Subsequently, the native promoter PppsA of the cluster responsible for the fengycin synthesis was replaced by the Pveg promoter, resulting in a further 5.22-fold increase in fengycin titer. To confer xylose utilization capacity to B. subtilis, deletion of araR and constitutive overexpression of araE were performed, and the xylose consumption rate of the engineered strain BSUY06 reached 0.29 g/L/h, which is about 6.25-fold higher than that of the parent strain BSUY04-1. In the final phase of this study, the fermentation characteristics were observed and the initial xylose concentration was optimized. In this study, 40 g/L xylose was proved to be the most suitable initial concentration for growth and fengycin fermentation, which leading to a fengycin titer of 430.86 mg/L. This study demonstrated that lignocellulose, the clean and sustainable substrate with xylose as the second largest sugar, is a potential substrate for the production of fengycin.

Predicting multifunctional peptides based on a multi-scale ResNet model combined with channel attention mechanisms

&lt;abstract&gt; &lt;p&gt;Peptides are biomolecules composed of multiple amino acid residues connected by peptide bonds, which are widely involved in physiological and biochemical processes in organisms and exhibit diverse functions. In previous studies, the focus was primarily on single-functional peptides. However, research trends indicate that an increasing number of multifunctional peptides are being identified and discovered. To address this challenge, we proposed a deep learning method based on multi-scale ResNet as the backbone combined with a channel attention mechanism (called MSRC) for the identification of multifunctional peptides. Furthermore, the data imbalance problem was solved through the comprehensive use of online data augmentation and confidence-based weighted loss functions. Experimental results demonstrated that the proposed MSRC method achieved an accuracy of 0.688 with an absolute true rate of 0.619. Notably, in predicting minority class peptides such as AEP, AHIVP, and BBP, the MSRC model exhibited heightened sensitivity, showcasing its exceptional capability in addressing issues related to minority classes. By enhancing the precision in identifying and predicting multifunctional peptides, the MSRC method was poised to contribute significantly to advancements in drug discovery, disease treatment, and biotechnology.&lt;/p&gt; &lt;/abstract&gt;

Delivery of rapamycin by biomimetic peptide nanoparticles targeting oxidized low-density lipoprotein in atherosclerotic plaques.

Drug delivery systems based on biomimetic peptide nanoparticles are steadily gaining prominence in the treatment of diverse medical conditions. This study focused on the development of peptides that depend on ligand-receptor interactions to load rapamycin (RAPA). Furthermore, a multifunctional peptide was engineered to target oxidized low-density lipoprotein (oxLDL) within atherosclerotic plaques, facilitating the localized delivery of RAPA. The interactions between peptides and RAPA/oxLDL were analyzed by simulations and experimental approaches. Results show that the main amino acid residues on the mammalian target of rapamycin that bind to RAPA are constructed as peptides (P1 and P2), which have specific interactions with RAPA and can effectively improve the loading efficiency of RAPA. The encapsulation and drug loading efficiencies of P1/P2 were 68.0/47.9% and 48.3/36.5%, respectively. In addition, the interaction force of the multifunctional peptide (P3) and oxLDL surpassed that of their interaction with human umbilical vein endothelial cells by a factor of 3.6, conclusively establishing the specific targeting of oxLDL by these nanoparticles. The encapsulation and drug loading efficiencies of P3 for RAPA were determined to be 60.2% and 41.5%. P3 can effectively load RAPA and target oxLDL within the plaque, suggesting that P3 has potential as a therapeutic agent for atherosclerotic disease.

Deep2Pep: A deep learning method in multi-label classification of bioactive peptide

Functional peptides are easy to absorb and have low side effects, which has attracted increasing interest from pharmaceutical scientists. However, due to the limitations in the laboratory funding and human resources, it is difficult to screen the functional peptides from a large number of peptides with unknown functions. With the development of machine learning and Deep learning, the combination of computational methods and biological information provides an effective method for identifying peptide functions. To explore the value of multi-functional active peptides, a new deep learning method named Deep2Pep (Deep learning to Peptides) was constructed, which was based on sequence encoding, embedding, and language tokenizer. It can achieve predictions of peptides on antimicrobial, antihypertensive, antioxidant and antihyperglycemic by converting sequence information into digital vectors, combined BiLSTM, attention-residual algorithm, and BERT Encoder. The results showed that Deep2Pep had a Hamming Loss of 0.095, subset Accuracy of 0.737, and Macro F1-Score of 0.734. which outperformed other models. BiLSTM played a primary role in Deep2Pep, which BERT encoder was in an auxiliary position. Deep learning algorithms was used in this study to accurately predict the four active functions of peptides, and it was expected to provide effective references for predicting multi-functional peptides.