As major components of spider venoms, neurotoxic peptides exhibit structural diversity, target specificity, and have great pharmaceutical potential. Deep learning may be an alternative to the laborious and time-consuming methods for identifying these peptides. However, the major hurdle in developing a deep learning model is the limited data on neurotoxic peptides. Here, we present a peptide data augmentation method that improves the recognition of neurotoxic peptides via a convolutional neural network model. The neurotoxic peptides were augmented with the known neurotoxic peptides from UniProt database, and the models were trained using a training set with or without the generated sequences to verify the augmented data. The model trained with the augmented dataset outperformed the one with the unaugmented dataset, achieving accuracy of 0.9953, precision of 0.9922, recall of 0.9984, and F1 score of 0.9953 in simulation dataset. From the set of all RNA transcripts of Callobius koreanus spider, we discovered neurotoxic peptides via the model, resulting in 275 putative peptides of which 252 novel sequences and only 23 sequences showing homology with the known peptides by Basic Local Alignment Search Tool. Among these 275 peptides, four were selected and shown to have neuromodulatory effects on the human neuroblastoma cell line SH-SY5Y. The augmentation method presented here may be applied to the identification of other functional peptides from biological resources with insufficient data.

Two bacterial strains, designated MJB4^T and SJ7^T, were isolated from water samples collected from Jeongbang Falls on Jeju Island, Republic of Korea. Phylogenetic analysis of 16S rRNA gene sequences indicated that the two strains belonged to the genera Nocardioides and Hyunsoonleella , owing to their high similarities to Nocardioides jensenii DSM 29641^T (97.5 %) and Hyunsoonleella rubra FA042 ^T (96.3 %), respectively. These values are much lower than the gold standard for bacterial species (98.7 %). The average nucleotide identity values between strains MJB4^T, SJ7^T and the reference strains, Nocardioides jensenii DSM 29641^T, Nocardioides daejeonensis MJ31^T and Hyunsoonleella flava T58^T were 77.2, 75.9 and 75.4 %, respectively. Strains MJB4^T and SJ7^T and the type strains of the species involved in system incidence have average nucleotide identity and average amino acid threshold values of 60.1–82.6 % for the species boundary (95–96 %), which confirms that strains MJB4^T and SJ7^T represent two new species of genus Nocardioides and Hyunsoonleella , respectively. Based on phylogenetic and phenotypic data, strains MJB4^T and SJ7^T are considered to represent novel species of the genus Nocardioides and Hyunsoonleella , respectively, for which the names Nocardioides donggukensis sp. nov. (type strain MJB4^T=KACC 21724^T=NBRC 114402^T) and Hyunsoonleella aquatilis sp. nov., (type strain SJ7^T=KACC 21715^T=NBRC 114486^T) have been proposed.

Dupuytren’s disease (DD) is a fibroproliferative disorder characterized by thickening and contracting palmar fascia. The exact pathogenesis of DD remains unknown. In this study, we identified co-expressed gene set (DD signature) consisting of 753 genes via weighted gene co-expression network analysis. To confirm the robustness of DD signature, module enrichment analysis and meta-analysis were performed. Moreover, this signature effectively classified DD disease samples. The DD signature were significantly enriched in unfolded protein response (UPR) related to endoplasmic reticulum (ER) stress. Next, we conducted multiple-phenotype regression analysis to identify trans-regulatory hotspots regulating expression levels of DD signature using Genotype-Tissue Expression data. Finally, 10 trans-regulatory hotspots and 16 eGenes genes that are significantly associated with at least one cis-eQTL were identified. Among these eGenes, major histocompatibility complex class II genes and ZFP57 zinc finger protein were closely related to ER stress and UPR, suggesting that these genetic markers might be potential therapeutic targets for DD.

Ion channels, which can be modulated by peptides, are promising drug targets for neurological, metabolic, and cardiovascular disorders. Because it is expensive and labor-intensive to experimentally screen ion channel-modulating peptides (IMPs), in-silico approaches can serve as excellent alternatives. In this study, we present PrIMP, prediction models for screening IMPs that can target sodium, potassium, and calcium ion channels, as well as nicotine acetylcholine receptors (nAChRs). To overcome the data insufficiency of the IMPs, we utilized two types of knowledge transfer approaches: multi-task learning (MTL) and transfer learning (TL). MTL enabled model training for four target tasks simultaneously with hard parameter sharing, thereby increasing model generalization. TL transferred knowledge of pre-trained model weights from antimicrobial peptide data, which was a much larger, naturally-occurring functional peptide dataset that could potentially improve the model performance. MTL and TL successfully improved the prediction performance of prediction models. In addition, a hybrid approach by implementing deep learning along with traditional machine learning was utilized, with additional performance improvements. PrIMP achieved F1 scores of 0.933 (sodium ion channel), 0.937 (potassium ion channel), 0.893 (calcium ion channel), and 0.931 (nAChRs). The pre-processed dataset and proposed model are available at https://github.com/bzlee-bio/PrIMP.

Background: The emergence and prevalence of antibiotic-resistant bacteria (ARBs) have become a serious global threat, as the morbidity and mortality associated with ARB infections are continuously rising. The activation of quorum sensing (QS) genes can promote biofilm formation, which contributes to the acquisition of drug resistance and increases virulence. Therefore, there is an urgent need to develop new antimicrobial agents to control ARB and prevent further development. Antimicrobial peptides (AMPs) are naturally occurring defense molecules in organisms known to suppress pathogens through a broad range of antimicrobial mechanisms. Methods: In this study, we utilized a previously developed deep-learning model to identify AMP candidates from the venom gland transcriptome of the spider Pardosa astrigera, followed by experimental validation. Results: PA-Win2 was among the top-scoring predicted peptides and was selected based on physiochemical features. Subsequent experimental validation demonstrated that PA-Win2 inhibits the growth of Bacillus subtilis, Escherichia coli, Staphylococcus aureus, Staphylococcus epidermidis, Pseudomonas aeruginosa, and multidrug-resistant P. aeruginosa (MRPA) strain CCARM 2095. The peptide exhibited strong bactericidal activity against P. aeruginosa, and MRPA CCARM 2095 through the depolarization of bacterial cytoplasmic membranes and alteration of gene expression associated with bacterial survival. In addition, PA-Win2 effectively inhibited biofilm formation and degraded pre-formed biofilms of P. aeruginosa. The gene expression study showed that the peptide treatment led to the downregulation of QS genes in the Las, Pqs, and Rhl systems. Conclusions: These findings suggest PA-Win2 as a promising drug candidate against ARB and demonstrate the potential of in silico methods in discovering functional peptides from biological data.