VPatho: a deep learning-based two-stage approach for accurate prediction of gain-of-function and loss-of-function variants

MetalTrans: A Biological Language Model-Based Approach for Predicting Disease-Associated Mutations in Protein Metal-Binding Sites

The critical importance of accurately predicting mutations in protein metal-binding sites for advancing drug discovery and enhancing disease diagnostic processes cannot be overstated. In response to this imperative, MetalTrans emerges as an accurate predictor for disease-associated mutations in protein metal-binding sites. The core innovation of MetalTrans lies in its seamless integration of multifeature splicing with the Transformer framework, a strategy that ensures exhaustive feature extraction. Central to MetalTrans's effectiveness is its deep feature combination strategy, which merges evolutionary-scale modeling amino acid embeddings with ProtTrans embeddings, thus shedding light on the biochemical properties of proteins. Employing the Transformer component, MetalTrans leverages the self-attention mechanism to delve into higher-level representations. Utilizing mutation site information for feature fusion not only enriches the feature set but also sidesteps the common pitfall of overestimation linked to protein sequence-based predictions. This nuanced approach to feature fusion is a key differentiator, enabling MetalTrans to outperform existing methods significantly, as evidenced by comparative analyses. Our evaluations across varied metal binding site data sets (specifically Zn, Ca, Mg, and Mix) underscore MetalTrans's superior performance, which achieved the average AUC values of 0.971, 0.965, 0.980, and 0.945 on multiple 5-fold cross-validation, respectively. Remarkably, against the multichannel convolutional neural network method on a benchmark independent test set, MetalTrans demonstrated unparalleled robustness and superiority, boasting the AUC score of 0.998 on multiple 5-fold cross-validation. Our comprehensive examination of the predicted outcomes further confirms the effectiveness of the model. The source codes, data sets, and prediction results for MetalTrans can be accessed for academic usage at https://github.com/EduardWang/MetalTrans.

PLMACPred prediction of anticancer peptides based on protein language model and wavelet denoising transformation

Anticancer peptides (ACPs) perform a promising role in discovering anti-cancer drugs. The growing research on ACPs as therapeutic agent is increasing due to its minimal side effects. However, identifying novel ACPs using wet-lab experiments are generally time-consuming, labor-intensive, and expensive. Leveraging computational methods for fast and accurate prediction of ACPs would harness the drug discovery process. Herein, a machine learning-based predictor, called PLMACPred, is developed for identifying ACPs from peptide sequence only. PLMACPred adopted a set of encoding schemes representing evolutionary-property, composition-property, and protein language model (PLM), i.e., evolutionary scale modeling (ESM-2)- and ProtT5-based embedding to encode peptides. Then, two-dimensional (2D) wavelet denoising (WD) was employed to remove the noise from extracted features. Finally, ensemble-based cascade deep forest (CDF) model was developed to identify ACP. PLMACPred model attained superior performance on all three benchmark datasets, namely, ACPmain, ACPAlter, and ACP740 over tenfold cross validation and independent dataset. PLMACPred outperformed the existing models and improved the prediction accuracy by 18.53%, 2.4%, 7.59% on ACPmain, ACPalter, ACP740 dataset, respectively. We showed that embedding from ProtT5 and ESM-2 was capable of capturing better contextual information from the entire sequence than the other encoding schemes for ACP prediction. For the explainability of proposed model, SHAP (SHapley Additive exPlanations) method was used to analyze the feature effect on the ACP prediction. A list of novel sequence motifs was proposed from the ACP sequence using MEME suites. We believe, PLMACPred will support in accelerating the discovery of novel ACPs as well as other activities of microbial peptides.

Distances from ligands as main predictive features for pathogenicity and functional effect of variants in NMDA receptors

Genetic variants in genes GRIN1 , GRIN2A , GRIN2B , and GRIN2D , which encode subunits of the N-methyl-D-aspartate receptor (NMDAR), have been associated with severe and heterogeneous neurologic diseases. Missense variants in these genes can result in gain or loss of the NMDAR function, requiring opposite therapeutic treatments. Computational methods that predict pathogenicity and molecular functional effects are therefore crucial for accurate diagnosis and therapeutic applications. We assembled missense variants: 201 from patients, 631 from general population, and 159 characterized by electrophysiological readouts showing whether they can enhance or reduce the receptor function. This includes new functional data from 47 variants reported here, for the first time. We found that pathogenic/benign variants and variants that increase/decrease the channel function were distributed unevenly on the protein structure, with spatial proximity to ligands bound to the agonist and antagonist binding sites being key predictive features. Leveraging distances from ligands, we developed two independent machine learning-based predictors for NMDAR missense variants: a pathogenicity predictor which outperforms currently available predictors (AUC=0.945, MCC=0.726), and the first binary predictor of molecular function (increase or decrease) (AUC=0.809, MCC=0.523). Using these, we reclassified variants of uncertain significance in the ClinVar database and refined a previous genome-informed epidemiological model to estimate the birth incidence of molecular mechanism-defined GRIN disorders. Our findings demonstrate that distance from ligands is an important feature in NMDARs that can enhance variant pathogenicity prediction and enable functional prediction. Further studies with larger numbers of phenotypically and functionally characterized variants will enhance the potential clinical utility of this method.

Review of Computational Methods and Database Sources for Predicting the Effects of Coding Frameshift Small Insertion and Deletion Variations

Genetic variations (including substitutions, insertions, and deletions) exert a profound influence on DNA sequences. These variations are systematically classified as synonymous, nonsynonymous, and nonsense, each manifesting distinct effects on proteins. The implementation of high-throughput sequencing has significantly augmented our comprehension of the intricate interplay between gene variations and protein structure and function, as well as their ramifications in the context of diseases. Frameshift variations, particularly small insertions and deletions (indels), disrupt protein coding and are instrumental in disease pathogenesis. This review presents a succinct review of computational methods, databases, current challenges, and future directions in predicting the consequences of coding frameshift small indels variations. We analyzed the predictive efficacy, reliability, and utilization of computational methods and variant account, reliability, and utilization of database. Besides, we also compared the prediction methodologies on GOF/LOF pathogenic variation data. Addressing the challenges pertaining to prediction accuracy and cross-species generalizability, nascent technologies such as AI and deep learning harbor immense potential to enhance predictive capabilities. The importance of interdisciplinary research and collaboration cannot be overstated for devising effective diagnosis, treatment, and prevention strategies concerning diseases associated with coding frameshift indels variations.

Genome-wide prediction of pathogenic gain- and loss-of-function variants from ensemble learning of a diverse feature set

Gain-of-function (GOF) variants give rise to increased/novel protein functions whereas loss-of-function (LOF) variants lead to diminished protein function. Experimental approaches for identifying GOF and LOF are generally slow and costly, whilst available computational methods have not been optimized to discriminate between GOF and LOF variants. We have developed LoGoFunc, a machine learning method for predicting pathogenic GOF, pathogenic LOF, and neutral genetic variants, trained on a broad range of gene-, protein-, and variant-level features describing diverse biological characteristics. LoGoFunc outperforms other tools trained solely to predict pathogenicity for identifying pathogenic GOF and LOF variants and is available at https://itanlab.shinyapps.io/goflof/ .

H2Opred: a robust and efficient hybrid deep learning model for predicting 2'-O-methylation sites in human RNA

2'-O-methylation (2OM) is the most common post-transcriptional modification of RNA. It plays a crucial role in RNA splicing, RNA stability and innate immunity. Despite advances in high-throughput detection, the chemical stability of 2OM makes it difficult to detect and map in messenger RNA. Therefore, bioinformatics tools have been developed using machine learning (ML) algorithms to identify 2OM sites. These tools have made significant progress, but their performances remain unsatisfactory and need further improvement. In this study, we introduced H2Opred, a novel hybrid deep learning (HDL) model for accurately identifying 2OM sites in human RNA. Notably, this is the first application of HDL in developing four nucleotide-specific models [adenine (A2OM), cytosine (C2OM), guanine (G2OM) and uracil (U2OM)] as well as a generic model (N2OM). H2Opred incorporated both stacked 1D convolutional neural network (1D-CNN) blocks and stacked attention-based bidirectional gated recurrent unit (Bi-GRU-Att) blocks. 1D-CNN blocks learned effective feature representations from 14 conventional descriptors, while Bi-GRU-Att blocks learned feature representations from five natural language processing-based embeddings extracted from RNA sequences. H2Opred integrated these feature representations to make the final prediction. Rigorous cross-validation analysis demonstrated that H2Opred consistently outperforms conventional ML-based single-feature models on five different datasets. Moreover, the generic model of H2Opred demonstrated a remarkable performance on both training and testing datasets, significantly outperforming the existing predictor and other four nucleotide-specific H2Opred models. To enhance accessibility and usability, we have deployed a user-friendly web server for H2Opred, accessible at https://balalab-skku.org/H2Opred/. This platform will serve as an invaluable tool for accurately predicting 2OM sites within human RNA, thereby facilitating broader applications in relevant research endeavors.