A proteome-scale map of the human interactome network.

Just as reference genome sequences revolutionized human genetics, reference maps of interactome networks will be critical to fully understand genotype-phenotype relationships. Here, we describe a systematic map of ?14,000 high-quality human binary protein-protein interactions. At equal quality, this map is ?30% larger than what is available from small-scale studies published in the literature in the last few decades. While currently available information is highly biased and only covers a relatively small portion of the proteome, our systematic map appears strikingly more homogeneous, revealing a "broader" human interactome network than currently appreciated. The map also uncovers significant interconnectivity between known and candidate cancer gene products, providing unbiased evidence for an expanded functional cancer landscape, while demonstrating how high-quality interactome models will help "connect the dots" of the genomic revolution.

Whole-genome sequencing in autism identifies hot spots for de novo germline mutation.

De novo mutation plays an important role in autism spectrum disorders (ASDs). Notably, pathogenic copy number variants (CNVs) are characterized by high mutation rates. We hypothesize that hypermutability is a property of ASD genes and may also include nucleotide-substitution hot spots. We investigated global patterns of germline mutation by whole-genome sequencing of monozygotic twins concordant for ASD and their parents. Mutation rates varied widely throughout the genome (by 100-fold) and could be explained by intrinsic characteristics of DNA sequence and chromatin structure. Dense clusters of mutations within individual genomes were attributable to compound mutation or gene conversion. Hypermutability was a characteristic of genes involved in ASD and other diseases. In addition, genes impacted by mutations in this study were associated with ASD in independent exome-sequencing data sets. Our findings suggest that regional hypermutation is a significant factor shaping patterns of genetic variation and disease risk in humans.

Genomic insights into genes expressed specifically during infancy highlight their dominant influence on the neuronal system.

BACKGROUND: Elucidating the dynamics of gene expression across developmental stages, including the genomic characteristics of brain expression during infancy, is pivotal in deciphering human psychiatric and neurological disorders and providing insights into developmental disorders. RESULTS: Leveraging comprehensive human GWAS associations with temporal and spatial brain expression data, we discovered a distinctive co-expression cluster comprising 897 genes highly expressed specifically during infancy, enriched in functions related to the neuronal system. This gene cluster notably harbors the highest ratio of genes linked to psychiatric and neurological disorders. Through computational analysis, MYT1L emerged as a potential central transcription factor governing these genes. Remarkably, the infancy-specific expressed genes, including SYT1, exhibit prominent colocalization within human accelerated regions. Additionally, chromatin state analysis unveiled prevalent epigenetic markers associated with enhancer-specific modifications. In addition, this cluster of genes has demonstrated to be specifically highly expressed in cell-types including excitatory neurons, medial ganglionic eminence and caudal ganglionic eminence. CONCLUSIONS: This study comprehensively characterizes the genomics and epigenomics of genes specifically expressed during infancy, identifying crucial hub genes and transcription factors. These findings offer valuable insights into early detection strategies and interventions for psychiatric and neurological disorders.

Locus-level antagonistic selection shaped the polygenic architecture of human complex diseases.

BACKGROUND: We aimed to evaluate the potential role of antagonistic selection in polygenic diseases: if one variant increases the risk of one disease and decreases the risk of another disease, the signals of genetic risk elimination by natural selection will be distorted, which leads to a higher frequency of risk alleles. METHODS: We applied local genetic correlations and transcriptome-wide association studies to identify genomic loci and genes adversely associated with at least two diseases. Then, we used different population genetic metrics to measure the signals of natural selection for these loci and genes. RESULTS: First, we identified 2120 cases of antagonistic pleiotropy (negative local genetic correlation) among 87 diseases in 716 genomic loci (antagonistic loci). Next, by comparing with non-antagonistic loci, we observed that antagonistic loci explained an excess proportion of disease heritability (median 6%), showed enhanced signals of balancing selection, and reduced signals of directional polygenic adaptation. Then, at the gene expression level, we identified 31,991 cases of antagonistic pleiotropy among 98 diseases at 4368 genes. However, evidence of altered signals of selection pressure and heritability distribution at the gene expression level is limited. CONCLUSION: We conclude that antagonistic pleiotropy is widespread among human polygenic diseases, and it has distorted the evolutionary signal and genetic architecture of diseases at the locus level.

Reduced motor cortex GABABR function following chronic alcohol exposure.

The GABAB receptor (GABABR) agonist baclofen has been used to treat alcohol and several other substance use disorders (AUD/SUD), yet its underlying neural mechanism remains unclear. The present study aimed to investigate cortical GABABR dynamics following chronic alcohol exposure. Ex vivo brain slice recordings from mice chronically exposed to alcohol revealed a reduction in GABABR-mediated currents, as well as a decrease of GABAB1/2R and G-protein-coupled inwardly rectifying potassium channel 2 (GIRK2) activities in the motor cortex. Moreover, our data indicated that these alterations could be attributed to dephosphorylation at the site of serine 783 (ser-783) in GABAB2 subunit, which regulates the surface expression of GABABR. Furthermore, a human study using paired-pulse-transcranial magnetic stimulation (TMS) analysis further demonstrated a reduced cortical inhibition mediated by GABABR in patients with AUD. Our findings provide the first evidence that chronic alcohol exposure is associated with significantly impaired cortical GABABR function. The ability to promote GABABR signaling may account for the therapeutic efficacy of baclofen in AUD.

Single-cell spatial transcriptomic analysis reveals common and divergent features of developing postnatal granule cerebellar cells and medulloblastoma.

BACKGROUND: Cerebellar neurogenesis involves the generation of large numbers of cerebellar granule neurons (GNs) throughout development of the cerebellum, a process that involves tight regulation of proliferation and differentiation of granule neuron progenitors (GNPs). A number of transcriptional regulators, including Math1, and the signaling molecules Wnt and Shh have been shown to have important roles in GNP proliferation and differentiation, and deregulation of granule cell development has been reported to be associated with the pathogenesis of medulloblastoma. While the progenitor/differentiation states of cerebellar granule cells have been broadly investigated, a more detailed association between developmental differentiation programs and spatial gene expression patterns, and how these lead to differential generation of distinct types of medulloblastoma remains poorly understood. Here, we provide a comparative single-cell spatial transcriptomics analysis to better understand the similarities and differences between developing granule and medulloblastoma cells. RESULTS: To acquire an enhanced understanding of the precise cellular states of developing cerebellar granule cells, we performed single-cell RNA sequencing of 24,919 murine cerebellar cells from granule neuron-specific reporter mice (Math1-GFP; Dcx-DsRed mice). Our single-cell analysis revealed that there are four major states of developing cerebellar granule cells, including two subsets of granule progenitors and two subsets of differentiating/differentiated granule neurons. Further spatial transcriptomics technology enabled visualization of their spatial locations in cerebellum. In addition, we performed single-cell RNA sequencing of 18,372 cells from Patched+/- mutant mice and found that the transformed granule cells in medulloblastoma closely resembled developing granule neurons of varying differentiation states. However, transformed granule neuron progenitors in medulloblastoma exhibit noticeably less tendency to differentiate compared with cells in normal development. CONCLUSION: In sum, our study revealed the cellular and spatial organization of the detailed states of cerebellar granule cells and provided direct evidence for the similarities and discrepancies between normal cerebellar development and tumorigenesis.

Loss of Setd2 associates with aberrant microRNA expression and contributes to inflammatory bowel disease progression in mice.

Both SETD2-mediated H3K36me3 and miRNAs play critical epigenetic roles in inflammatory bowel disease (IBD) and involve in the dysfunctional intestinal barrier. However, little is known about cross-talk between these two types of regulators in IBD progression. We performed small RNA sequencing of Setd2 epithelium-specific knockout mice (Setd2Vil-KO) and wild-type controls, both with DSS-induced colitis, and designed a framework for integrative analysis. Firstly, we integrated the downloaded ChIP-seq data with miRNA expression profiles and identified a significant intersection of pre-miRNA expression and H3K36me3 modification. A significant inverse correlation was detected between changes of H3K36me3 modification and expression of the 171 peak-covered miRNAs. We further integrated RNA-seq data with predicted miRNA targets to screen negatively regulated miRNA-mRNA pairs and found the H3K36me3-associated differentially expressed microRNAs significantly enriched in cell-cell junction and signaling pathways. Using network analysis, we identified ten hub miRNAs, among which six are H3K36me3-associated, suggesting therapeutic targets for IBD patients with SETD2-deficiency.

An Integrative Analysis of Identified Schizophrenia-Associated Brain Cell Types and Gene Expression Changes.

Schizophrenia (SCZ) is a severe mental disorder that may result in hallucinations, delusions, and extremely disordered thinking. How each cell type in the brain contributes to SCZ occurrence is still unclear. Here, we leveraged the human dorsolateral prefrontal cortex bulk RNA-seq data, then used the RNA-seq deconvolution algorithm CIBERSORTx to generate SCZ brain single-cell RNA-seq data for a comprehensive analysis to understand SCZ-associated brain cell types and gene expression changes. Firstly, we observed that the proportions of brain cell types in SCZ differed from normal samples. Among these cell types, astrocyte, pericyte, and PAX6 cells were found to have a higher proportion in SCZ patients (astrocyte: SCZ = 0.163, control = 0.145, P.adj = 4.9 × 10-4, effect size = 0.478; pericyte: SCZ = 0.057, control = 0.066, P.adj = 1.1 × 10-4, effect size = 0.519; PAX6: SCZ = 0.014, control = 0.011, P.adj = 0.014, effect size = 0.377), while the L5/6_IT_CAR3 cells and LAMP5 cells are the exact opposite (L5/6_IT_Car3: SCZ = 0.102, control = 0.108, P.adj = 0.016, effect size = 0.369; LAMP5: SCZ = 0.057, control = 0.066, P.adj = 2.2 × 10-6, effect size = 0.617). Next, we investigated gene expression in cell types and functional pathways in SCZ. We observed chemical synaptic transmission dysregulation in two types of GABAergic neurons (PVALB and LAMP5), and immune reaction involvement in GABAergic neurons (SST) and non-neuronal cell types (endothelial and oligodendrocyte). Furthermore, we observed that some differential expression genes from bulk RNA-seq displayed cell-type-specific abnormalities in the expression of molecules in SCZ. Finally, the cell types with the SCZ-related transcriptomic changes could be considered to belong to the same module since we observed two major similar coordinated transcriptomic changes across these cell types. Together, our results offer novel insights into cellular heterogeneity and the molecular mechanisms underlying SCZ.

Mendelian randomization studies of brain MRI yield insights into the pathogenesis of neuropsychiatric disorders.

BACKGROUND: Observational studies have identified various associations between neuroimaging alterations and neuropsychiatric disorders. However, whether such associations could truly reflect causal relations remains still unknown. RESULTS: Here, we leveraged genome-wide association studies (GWAS) summary statistics for (1) 11 psychiatric disorders (sample sizes varied from n = 9,725 to 1,331,010); (2) 110 diffusion tensor imaging (DTI) measurement (sample size n = 17,706); (3) 101 region-of-interest (ROI) volumes, and investigate the causal relationship between brain structures and neuropsychiatric disorders by two-sample Mendelian randomization. Among all DTI-Disorder combinations, we observed a significant causal association between the superior longitudinal fasciculus (SLF) and the risk of Anorexia nervosa (AN) (Odds Ratio [OR] = 0.62, 95 % confidence interval: 0.50 ~ 0.76, P = 6.4 × 10- 6). Similar significant associations were also observed between the body of the corpus callosum (fractional anisotropy) and Alzheimer's disease (OR = 1.07, 95 % CI: 1.03 ~ 1.11, P = 4.1 × 10- 5). By combining all observations, we found that the overall p-value for DTI - Disorder associations was significantly elevated compared to the null distribution (Kolmogorov-Smirnov P = 0.009, inflation factor λ = 1.37), especially for DTI - Bipolar disorder (BP) (λ = 2.64) and DTI - AN (λ = 1.82). In contrast, for ROI-Disorder combinations, we only found a significant association between the brain region of pars triangularis and Schizophrenia (OR = 0.48, 95 % CI: 0.34 ~ 0.69, P = 5.9 × 10- 5) and no overall p-value elevation for ROI-Disorder analysis compared to the null expectation. CONCLUSIONS: As a whole, we show that SLF degeneration may be a risk factor for AN, while DTI variations could be causally related to some neuropsychiatric disorders, such as BP and AN. Also, the white matter structure might have a larger impact on neuropsychiatric disorders than subregion volumes.

Single-cell RNA sequencing reveals the epithelial cell heterogeneity and invasive subpopulation in human bladder cancer.

Bladder cancer represents a highly heterogeneous disease characterized by distinct histological, molecular and clinical phenotypes, and a detailed analysis of tumor cell invasion and crosstalks within bladder tumor cells has not been determined. Here, we applied droplet-based single-cell RNA sequencing (scRNA-seq) to acquire transcriptional profiles of 36 619 single cells isolated from seven patients. Single cell transcriptional profiles matched well with the pathological basal/luminal subtypes. Notably, in T1 tumors diagnosed as luminal subtype, basal cells displayed characteristics of epithelial-mesenchymal transition (EMT) and mainly located at the tumor-stromal interface as well as micrometastases in the lamina propria. In one T3 tumor, muscle-invasive tumor showed significantly higher expression of cancer stem cell markers SOX9 and SOX2 than the primary tumor. We additionally analyzed communications between tumor cells and demonstrated its relevance to basal/luminal phenotypes. Overall, our single-cell study provides a deeper insight into the tumor cell heterogeneity associated with bladder cancer progression.

Frequency and Complexity of De Novo Structural Mutation in Autism.

Genetic studies of autism spectrum disorder (ASD) have established that de novo duplications and deletions contribute to risk. However, ascertainment of structural variants (SVs) has been restricted by the coarse resolution of current approaches. By applying a custom pipeline for SV discovery, genotyping, and de novo assembly to genome sequencing of 235 subjects (71 affected individuals, 26 healthy siblings, and their parents), we compiled an atlas of 29,719 SV loci (5,213/genome), comprising 11 different classes. We found a high diversity of de novo mutations, the majority of which were undetectable by previous methods. In addition, we observed complex mutation clusters where combinations of de novo SVs, nucleotide substitutions, and indels occurred as a single event. We estimate a high rate of structural mutation in humans (20%) and propose that genetic risk for ASD is attributable to an elevated frequency of gene-disrupting de novo SVs, but not an elevated rate of genome rearrangement.

Expression alteration of microRNAs in Nucleus Accumbens is associated with chronic stress and antidepressant treatment in rats.

BACKGROUND: Nucleus Accumbens (NAc) is a vital brain region for the process of reward and stress, whereas microRNA plays a crucial role in depression pathology. However, the abnormality of NAc miRNA expression during the stress-induced depression and antidepressant treatment, as well as its biological significance, are still unknown. METHODS: We performed the small RNA-sequencing in NAc of rats from three groups: control, chronic unpredictable mild stress (CUMS), and CUMS with an antidepressant, Escitalopram. We applied an integrative pipeline for analyzing the miRNA expression alternation in different model groups, including differential expression analysis, co-expression analysis, as well as a subsequent pathway/network analysis to discover both miRNA alteration pattern and its biological significance. RESULT: A total of 423 miRNAs were included in analysis.18/8 differential expressing (DE) miRNA (adjusted p < 0.05, |log2FC| > 1) were observed in controls Vs. depression/depression Vs. treatment, 2 of which are overlapping. 78% (14/18) of these miRNAs showed opposite trends of alteration in stress and treatment. Two micro RNA, miR-10b-5p and miR-214-3p, appeared to be hubs in the regulation networks and also among the top findings in both differential analyses. Using co-expression analysis, we found a functional module that strongly correlated with stress (R = 0.96, P = 0.003), and another functional module with a moderate correlation with anhedonia (R = 0.89, P = 0.02). We also found that predicted targets of these miRNAs were significantly enriched in the Ras signaling pathway, which is associated with both depression, anhedonia, and antidepressant treatment. CONCLUSION: Escitalopram treatment can significantly reverse NAc miRNA abnormality induced by chronic stress. However, the novel miRNA alteration that is absent in stress pathology also emerges, which means that antidepressant treatment is unlikely to bring miRNA expression back to the same level as the controls. Also, the Ras-signaling pathway may be involved in explaining the depression disease etiology, the clinical symptom, and treatment response of stress-induced depression.

When loss-of-function is loss of function: assessing mutational signatures and impact of loss-of-function genetic variants.

MOTIVATION: Loss-of-function genetic variants are frequently associated with severe clinical phenotypes, yet many are present in the genomes of healthy individuals. The available methods to assess the impact of these variants rely primarily upon evolutionary conservation with little to no consideration of the structural and functional implications for the protein. They further do not provide information to the user regarding specific molecular alterations potentially causative of disease. RESULTS: To address this, we investigate protein features underlying loss-of-function genetic variation and develop a machine learning method, MutPred-LOF, for the discrimination of pathogenic and tolerated variants that can also generate hypotheses on specific molecular events disrupted by the variant. We investigate a large set of human variants derived from the Human Gene Mutation Database, ClinVar and the Exome Aggregation Consortium. Our prediction method shows an area under the Receiver Operating Characteristic curve of 0.85 for all loss-of-function variants and 0.75 for proteins in which both pathogenic and neutral variants have been observed. We applied MutPred-LOF to a set of 1142 de novo vari3ants from neurodevelopmental disorders and find enrichment of pathogenic variants in affected individuals. Overall, our results highlight the potential of computational tools to elucidate causal mechanisms underlying loss of protein function in loss-of-function variants. AVAILABILITY AND IMPLEMENTATION: http://mutpred.mutdb.org. CONTACT: predrag@indiana.edu.

Haplotype function score improves biological interpretation and cross-ancestry polygenic prediction of human complex traits.

We propose a new framework for human genetic association studies: at each locus, a deep learning model (in this study, Sei) is used to calculate the functional genomic activity score for two haplotypes per individual. This score, defined as the Haplotype Function Score (HFS), replaces the original genotype in association studies. Applying the HFS framework to 14 complex traits in the UK Biobank, we identified 3619 independent HFS-trait associations with a significance of p < 5 × 10-8. Fine-mapping revealed 2699 causal associations, corresponding to a median increase of 63 causal findings per trait compared with single-nucleotide polymorphism (SNP)-based analysis. HFS-based enrichment analysis uncovered 727 pathway-trait associations and 153 tissue-trait associations with strong biological interpretability, including 'circadian pathway-chronotype' and 'arachidonic acid-intelligence'. Lastly, we applied least absolute shrinkage and selection operator (LASSO) regression to integrate HFS prediction score with SNP-based polygenic risk scores, which showed an improvement of 16.1-39.8% in cross-ancestry polygenic prediction. We concluded that HFS is a promising strategy for understanding the genetic basis of human complex traits.

Scattered throughout the human genome are variations in the genetic code that make individuals more or less likely to develop certain traits. To identify these variants, scientists carry out Genome-wide association studies (GWAS) which compare the DNA variants of large groups of people with and without the trait of interest. This method has been able to find the underlying genes for many human diseases, but it has limitations. For instance, some variations are linked together due to where they are positioned within DNA, which can result in GWAS falsely reporting associations between genetic variants and traits. This phenomenon, known as linkage equilibrium, can be avoided by analyzing functional genomics which looks at the multiple ways a gene's activity can be influenced by a variation. For instance, how the gene is copied and decoded in to proteins and RNA molecules, and the rate at which these products are generated. Researchers can now use an artificial intelligence technique called deep learning to generate functional genomic data from a particular DNA sequence. Here, Song et al. used one of these deep learning models to calculate the functional genomics of haplotypes, groups of genetic variants inherited from one parent. The approach was applied to DNA samples from over 350 thousand individuals included in the UK BioBank. An activity score, defined as the haplotype function score (or HFS for short), was calculated for at least two haplotypes per individual, and then compared to various complex traits like height or bone density. Song et al. found that the HFS framework was better at finding links between genes and specific traits than existing methods. It also provided more information on the biology that may be underpinning these outcomes. Although more work is needed to reduce the computer processing times required to calculate the HFS, Song et al. believe that their new method has the potential to improve the way researchers identify links between genes and human traits.

Human lineage mutations regulate RNA-protein binding of conserved genes NTRK2 and ITPR1 involved in human evolution.

Background: The role of human lineage mutations (HLMs) in human evolution through post-transcriptional modification is unclear. Aims: To investigate the contribution of HLMs to human evolution through post-transcriptional modification. Methods: We applied a deep learning model Seqweaver to predict how HLMs impact RNA-binding protein affinity. Results: We found that only 0.27% of HLMs had significant impacts on RNA-binding proteins at the threshold of the top 1% of human common variations. These HLMs enriched in a set of conserved genes highly expressed in adult excitatory neurons and prenatal Purkinje neurons, and were involved in synapse organisation and the GTPase pathway. These genes also carried excess damaging coding mutations that caused neurodevelopmental disorders, ataxia and schizophrenia. Among these genes, NTRK2 and ITPR1 had the most aggregated evidence of functional importance, suggesting their essential roles in cognition and bipedalism. Conclusions: Our findings suggest that a small subset of human-specific mutations have contributed to human speciation through impacts on post-transcriptional modification of critical brain-related genes.

Causality Investigation between Gut Microbiome and Sleep-Related Traits: A Bidirectional Two-Sample Mendelian Randomization Study.

Recent research has highlighted associations between sleep and microbial taxa and pathways. However, the causal effect of these associations remains unknown. To investigate this, we performed a bidirectional two-sample Mendelian randomization (MR) analysis using summary statistics of genome-wide association studies (GWAS) from 412 gut microbiome traits (N = 7738) and GWAS studies from seven sleep-associated traits (N = 345,552 to 386,577). We employed multiple MR methods to assess causality, with Inverse Variance Weighted (IVW) as the primary method, alongside a Bonferroni correction ((p < 2.4 × 10-4) to determine significant causal associations. We further applied Cochran's Q statistical analysis, MR-Egger intercept, and Mendelian randomization pleiotropy residual sum and outlier (MR-PRESSO) for heterogeneity and pleiotropy assessment. IVW estimates revealed 79 potential causal effects of microbial taxa and pathways on sleep-related traits and 45 inverse causal relationships, with over half related to pathways, emphasizing their significance. The results revealed two significant causal associations: genetically determined relative abundance of pentose phosphate decreased sleep duration (p = 9.00 × 10-5), and genetically determined increase in fatty acid level increased the ease of getting up in the morning (p = 8.06 × 10-5). Sensitivity analyses, including heterogeneity and pleiotropy tests, as well as a leave-one-out analysis of single nucleotide polymorphisms, confirmed the robustness of these relationships. This study explores the potential causal relationships between sleep and microbial taxa and pathways, offering novel insights into their complex interplay.

Interaction between RNF4 and SART3 is associated with the risk of schizophrenia.

The pathogenesis of schizophrenia (SCZ) is heavily influenced by genetic factors. Ring finger protein 4 (RNF4) and squamous cell carcinoma antigen recognized by T cells 3 (SART3) are thought to be involved in nervous system growth and development via oxidative stress pathways. Moreover, they have previously been linked to SCZ. Yet the role of RNF4 and SART3 in SCZ remains unclear. Here, we investigated how these two genes are involved in SCZ by studying their variants observed in patients. We first observed significantly elevated mRNA levels of RNF4 and SART3 in the peripheral blood in both first-episode (n = 30) and chronic (n = 30) SCZ patients compared to controls (n = 60). Next, we targeted-sequenced three single nucleotide polymorphisms (SNPs) in SART3 and six SNPs in RNF4 for association with SCZ using the genomic DNA extracted from peripheral blood leukocytes from SCZ participants (n = 392) and controls (n = 572). We observed a combination of SNPs that included rs1203860, rs2282765 (both in RNF4), and rs2287550 (in SART3) was associated with increased risk of SCZ, suggesting common pathogenic mechanisms between these two genes. We then conducted experiments in HEK293T cells to better understand the interaction between RNF4 and SART3. We observed that SART3 lowered the expression of RNF4 through ubiquitination and downregulated the expression of nuclear factor E2-related factor 2 (NRF2), a downstream factor of RNF4, implicating the existence of a possible shared regulatory mechanism for RNF4 and SART3. In conclusion, our study provides evidence that the interaction between RNF4 and SART3 contributes to the risk of SCZ. The findings shed light on the underlying molecular mechanisms of SCZ and may lead to the development of new therapies and interventions for this disorder.

Emden: A novel method integrating graph and transformer representations for predicting the effect of mutations on clinical drug response.

Precision medicine based on personalized genomics provides promising strategies to enhance the efficacy of molecular-targeted therapies. However, the clinical effectiveness of drugs has been severely limited due to genetic variations that lead to drug resistance. Predicting the impact of missense mutations on clinical drug response is an essential way to reduce the cost of clinical trials and understand genetic diseases. Here, we present Emden, a novel method integrating graph and transformer representations that predicts the effect of missense mutations on drug response through binary classification with interpretability. Emden utilized protein sequences-based features and drug structures as inputs for rapid prediction, employing competitive representation learning and demonstrating strong generalization capabilities and robustness. Our study showed promising potential for clinical drug guidance and deep insight into computer-assisted precision medicine. Emden is freely available as a web server at https://www.psymukb.net/Emden.

Inferring the Effects of Protein Variants on Protein-Protein Interactions with Interpretable Transformer Representations.

Identifying pathogenetic variants and inferring their impact on protein-protein interactions sheds light on their functional consequences on diseases. Limited by the availability of experimental data on the consequences of protein interaction, most existing methods focus on building models to predict changes in protein binding affinity. Here, we introduced MIPPI, an end-to-end, interpretable transformer-based deep learning model that learns features directly from sequences by leveraging the interaction data from IMEx. MIPPI was specifically trained to determine the types of variant impact (increasing, decreasing, disrupting, and no effect) on protein-protein interactions. We demonstrate the accuracy of MIPPI and provide interpretation through the analysis of learned attention weights, which exhibit correlations with the amino acids interacting with the variant. Moreover, we showed the practicality of MIPPI in prioritizing de novo mutations associated with complex neurodevelopmental disorders and the potential to determine the pathogenic and driving mutations. Finally, we experimentally validated the functional impact of several variants identified in patients with such disorders. Overall, MIPPI emerges as a versatile, robust, and interpretable model, capable of effectively predicting mutation impacts on protein-protein interactions and facilitating the discovery of clinically actionable variants.