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: Regulatory hotspots are genetic variations that may regulate the expression levels of many genes. It has been of great interest to find those hotspots utilizing expression quantitative trait locus (eQTL) analysis. However, it has been reported that many of the findings are spurious hotspots induced by various unknown confounding factors. Recently, methods utilizing complicated statistical models have been developed that successfully identify genuine hotspots. Next-generation Intersample Correlation Emended (NICE) is one of the methods that show high sensitivity and low false-discovery rate in finding regulatory hotspots. Even though the methods successfully find genuine hotspots, they have not been widely used due to their non-user-friendly interfaces and complex running processes. Furthermore, most of the methods are impractical due to their prohibitively high computational complexity. Results: To overcome the limitations of existing methods, we developed a fully automated web-based tool, referred to as NICER (NICE Renew), which is based on NICE program. First, we dramatically reduced running and installing burden of NICE. Second, we significantly reduced running time by incorporating multi-processing. Third, besides our web-based NICER, users can use NICER on Google Compute Engine and can readily install and run the NICER web service on their local computers. Finally, we provide different input formats and visualizations tools to show results. Utilizing a yeast dataset, we show that NICER can be successfully used in an eQTL analysis to identify many genuine regulatory hotspots, for which more than half of the hotspots were previously reported elsewhere. Conclusions: Even though many hotspot analysis tools have been proposed, they have not been widely used for many practical reasons. NICER is a fully-automated web-based solution for eQTL mapping and regulatory hotspots analysis. NICER provides a user-friendly interface and has made hotspot analysis more viable by reducing the running time significantly. We believe that NICER will become the method of choice for increasing power of eQTL hotspot analysis.

GWASs have successfully identified numerous genetic variants linked to complex traits, but traditional univariate approaches often fail to capture shared genetic architecture across multiple phenotypes. As the scale of genomic data continues to increase, the demand for more efficient multi-phenotype analysis methods has become particularly critical. In addition, the issue of population structure must also be properly addressed to ensure robust and unbiased results. Multivariate methods for multi-phenotype analysis, such as GAMMA, address this by combining linear mixed models with multivariate distance matrix regression to account for population structure; however, since these methods utilize computationally intensive models, developing efficient implementations is essential for practical analysis. Although GAMMA is a well-designed and effective tool, its original implementation relies on multiple programming environments and requires frequent data exchanges between components. These factors increase computational burden and complicate installation and execution for users unfamiliar with programming, making practical applications, particularly for high-dimensional datasets, challenging. Here, we present GAMMA-RAY, a high-performance C++ implementation that streamlines the computational pipeline, leverages parallel processing, and employs efficient matrix operations to achieve substantial reductions in runtime and memory usage. GAMMA-RAY provides both a user-friendly web-based interface for non-programmers and a standalone version for secure local execution. We further applied GAMMA-RAY to a yeast dataset and identified putative trans-eQTLs, in which several variants overlapped with previously reported cis- and trans-eQTLs. In addition, functional enrichment analysis revealed that the associated trans-eGenes are enriched, a conclusion consistently supported by biological annotation resources and underscoring the biological significance of these results.