AttOmics: attention-based architecture for diagnosis and prognosis from omics data.

MOTIVATION: The increasing availability of high-throughput omics data allows for considering a new medicine centered on individual patients. Precision medicine relies on exploiting these high-throughput data with machine-learning models, especially the ones based on deep-learning approaches, to improve diagnosis. Due to the high-dimensional small-sample nature of omics data, current deep-learning models end up with many parameters and have to be fitted with a limited training set. Furthermore, interactions between molecular entities inside an omics profile are not patient specific but are the same for all patients. RESULTS: In this article, we propose AttOmics, a new deep-learning architecture based on the self-attention mechanism. First, we decompose each omics profile into a set of groups, where each group contains related features. Then, by applying the self-attention mechanism to the set of groups, we can capture the different interactions specific to a patient. The results of different experiments carried out in this article show that our model can accurately predict the phenotype of a patient with fewer parameters than deep neural networks. Visualizing the attention maps can provide new insights into the essential groups for a particular phenotype. AVAILABILITY AND IMPLEMENTATION: The code and data are available at https://forge.ibisc.univ-evry.fr/abeaude/AttOmics. TCGA data can be downloaded from the Genomic Data Commons Data Portal.

A2Sign: Agnostic Algorithms for Signatures-a universal method for identifying molecular signatures from transcriptomic datasets prior to cell-type deconvolution.

MOTIVATION: Molecular signatures are critical for inferring the proportions of cell types from bulk transcriptomics data. However, the identification of these signatures is based on a methodology that relies on prior biological knowledge of the cell types being studied. When working with less known biological material, a data-driven approach is required to uncover the underlying classes and generate ad hoc signatures from healthy or pathogenic tissue. RESULTS: We present a new approach, A2Sign: Agnostic Algorithms for Signatures, based on a non-negative tensor factorization (NTF) strategy that allows us to identify cell-type-specific molecular signatures, greatly reduce collinearities and also account for inter-individual variability. We propose a global framework that can be applied to uncover molecular signatures for cell-type deconvolution in arbitrary tissues using bulk transcriptome data. We also present two new molecular signatures for deconvolution of up to 16 immune cell types using microarray or RNA-seq data. AVAILABILITY AND IMPLEMENTATION: All steps of our analysis were implemented in annotated Python notebooks (https://github.com/paulfogel/A2SIGN). To perform NTF, we used the NMTF package, which can be downloaded using Python pip install. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Structural basis for substrate selectivity and nucleophilic substitution mechanisms in human adenine phosphoribosyltransferase catalyzed reaction.

The reversible adenine phosphoribosyltransferase enzyme (APRT) is essential for purine homeostasis in prokaryotes and eukaryotes. In humans, APRT (hAPRT) is the only enzyme known to produce AMP in cells from dietary adenine. APRT can also process adenine analogs, which are involved in plant development or neuronal homeostasis. However, the molecular mechanism underlying substrate specificity of APRT and catalysis in both directions of the reaction remains poorly understood. Here we present the crystal structures of hAPRT complexed to three cellular nucleotide analogs (hypoxanthine, IMP, and GMP) that we compare with the phosphate-bound enzyme. We established that binding to hAPRT is substrate shape-specific in the forward reaction, whereas it is base-specific in the reverse reaction. Furthermore, a quantum mechanics/molecular mechanics (QM/MM) analysis suggests that the forward reaction is mainly a nucleophilic substitution of type 2 (SN2) with a mix of SN1-type molecular mechanism. Based on our structural analysis, a magnesium-assisted SN2-type mechanism would be involved in the reverse reaction. These results provide a framework for understanding the molecular mechanism and substrate discrimination in both directions by APRTs. This knowledge can play an instrumental role in the design of inhibitors, such as antiparasitic agents, or adenine-based substrates.

Performance of epistasis detection methods in semi-simulated GWAS.

BACKGROUND: Part of the missing heritability in Genome Wide Association Studies (GWAS) is expected to be explained by interactions between genetic variants, also called epistasis. Various statistical methods have been developed to detect epistasis in case-control GWAS. These methods face major statistical challenges due to the number of tests required, the complexity of the Linkage Disequilibrium (LD) structure, and the lack of consensus regarding the definition of epistasis. Their limited impact in terms of uncovering new biological knowledge might be explained in part by the limited amount of experimental data available to validate their statistical performances in a realistic GWAS context. In this paper, we introduce a simulation pipeline for generating real scale GWAS data, including epistasis and realistic LD structure. We evaluate five exhaustive bivariate interaction methods, fastepi, GBOOST, SHEsisEpi, DSS, and IndOR. Two hundred thirty four different disease scenarios are considered in extensive simulations. We report the performances of each method in terms of false positive rate control, power, area under the ROC curve (AUC), and computation time using a GPU. Finally we compare the result of each methods on a real GWAS of type 2 diabetes from the Welcome Trust Case Control Consortium. RESULTS: GBOOST, SHEsisEpi and DSS allow a satisfactory control of the false positive rate. fastepi and IndOR present an increase in false positive rate in presence of LD between causal SNPs, with our definition of epistasis. DSS performs best in terms of power and AUC in most scenarios with no or weak LD between causal SNPs. All methods can exhaustively analyze a GWAS with 6.105 SNPs and 15,000 samples in a couple of hours using a GPU. CONCLUSION: This study confirms that computation time is no longer a limiting factor for performing an exhaustive search of epistasis in large GWAS. For this task, using DSS on SNP pairs with limited LD seems to be a good strategy to achieve the best statistical performance. A combination approach using both DSS and GBOOST is supported by the simulation results and the analysis of the WTCCC dataset demonstrated that this approach can detect distinct genes in epistasis. Finally, weak epistasis between common variants will be detectable with existing methods when GWAS of a few tens of thousands cases and controls are available.

Large-scale computational modelling of the M1 and M2 synovial macrophages in rheumatoid arthritis.

Macrophages play an essential role in rheumatoid arthritis. Depending on their phenotype (M1 or M2), they can play a role in the initiation or resolution of inflammation. The M1/M2 ratio in rheumatoid arthritis is higher than in healthy controls. Despite this, no treatment targeting specifically macrophages is currently used in clinics. Thus, devising strategies to selectively deplete proinflammatory macrophages and promote anti-inflammatory macrophages could be a promising therapeutic approach. State-of-the-art molecular interaction maps of M1 and M2 macrophages in rheumatoid arthritis are available and represent a dense source of knowledge; however, these maps remain limited by their static nature. Discrete dynamic modelling can be employed to study the emergent behaviours of these systems. Nevertheless, handling such large-scale models is challenging. Due to their massive size, it is computationally demanding to identify biologically relevant states in a cell- and disease-specific context. In this work, we developed an efficient computational framework that converts molecular interaction maps into Boolean models using the CaSQ tool. Next, we used a newly developed version of the BMA tool deployed to a high-performance computing cluster to identify the models' steady states. The identified attractors are then validated using gene expression data sets and prior knowledge. We successfully applied our framework to generate and calibrate the M1 and M2 macrophage Boolean models for rheumatoid arthritis. Using KO simulations, we identified NFkB, JAK1/JAK2, and ERK1/Notch1 as potential targets that could selectively suppress proinflammatory macrophages and GSK3B as a promising target that could promote anti-inflammatory macrophages in rheumatoid arthritis.

Phenotypic variation among seven members of one family with deficiency of hypoxanthine-guanine phosphoribosyltransferase.

We describe a family of seven boys affected by Lesch-Nyhan disease with various phenotypes. Further investigations revealed a mutation c.203T>C in the gene encoding HGprt of all members, with substitution of leucine to proline at residue 68 (p.Leu68Pro). Thus patients from this family display a wide variety of symptoms although sharing the same mutation. Mutant HGprt enzyme was prepared by site-directed mutagenesis and the kinetics of the enzyme revealed that the catalytic activity of the mutant was reduced, in association with marked reductions in the affinity towards phosphoribosylpyrophosphate (PRPP). Its Km for PRPP was increased 215-fold with hypoxanthine as substrate and 40-fold with guanine as substrate with associated reduced catalytic potential. Molecular modeling confirmed that the most prominent defect was the dramatically reduced affinity towards PRPP. Our studies suggest that the p.Leu68Pro mutation has a strong impact on PRPP binding and on stability of the active conformation. This suggests that factors other than HGprt activity per se may influence the phenotype of Lesch-Nyhan patients.

DiSiR: fast and robust method to identify ligand-receptor interactions at subunit level from single-cell RNA-sequencing data.

Most cell-cell interactions and crosstalks are mediated by ligand-receptor interactions. The advent of single-cell RNA-sequencing (scRNA-seq) techniques has enabled characterizing tissue heterogeneity at single-cell level. In the past few years, several methods have been developed to study ligand-receptor interactions at cell type level using scRNA-seq data. However, there is still no easy way to query the activity of a specific user-defined signaling pathway in a targeted way or to map the interactions of the same subunit with different ligands as part of different receptor complexes. Here, we present DiSiR, a fast and easy-to-use permutation-based software framework to investigate how individual cells are interacting with each other by analyzing signaling pathways of multi-subunit ligand-activated receptors from scRNA-seq data, not only for available curated databases of ligand-receptor interactions, but also for interactions that are not listed in these databases. We show that, when utilized to infer ligand-receptor interactions from both simulated and real datasets, DiSiR outperforms other well-known permutation-based methods, e.g. CellPhoneDB and ICELLNET. Finally, to demonstrate DiSiR's utility in exploring data and generating biologically relevant hypotheses, we apply it to COVID lung and rheumatoid arthritis (RA) synovium scRNA-seq datasets and highlight potential differences between inflammatory pathways at cell type level for control versus disease samples.

Single-cell RNA-sequencing of PBMCs from SAVI patients reveals disease-associated monocytes with elevated integrated stress response.

Gain-of-function mutations in stimulator of interferon gene 1 (STING1) result in STING-associated vasculopathy with onset in infancy (SAVI), a severe autoinflammatory disease. Although elevated type I interferon (IFN) production is thought to be the leading cause of the symptoms observed in patients, STING can induce a set of pathways, which have roles in the onset and severity of SAVI and remain to be elucidated. To this end, we performed a multi-omics comparative analysis of peripheral blood mononuclear cells (PBMCs) and plasma from SAVI patients and healthy controls, combined with a dataset of healthy PBMCs treated with IFN-ß. Our data reveal a subset of disease-associated monocyte, expressing elevated CCL3, CCL4, and IL-6, as well as a strong integrated stress response, which we suggest is the result of direct PERK activation by STING. Cell-to-cell communication inference indicates that these monocytes lead to T cell early activation, resulting in their senescence and apoptosis. Last, we propose a transcriptomic signature of STING activation, independent of type I IFN response.

Benchmarking computational methods for B-cell receptor reconstruction from single-cell RNA-seq data.

Multiple methods have recently been developed to reconstruct full-length B-cell receptors (BCRs) from single-cell RNA sequencing (scRNA-seq) data. This need emerged from the expansion of scRNA-seq techniques, the increasing interest in antibody-based drug development and the importance of BCR repertoire changes in cancer and autoimmune disease progression. However, a comprehensive assessment of performance-influencing factors such as the sequencing depth, read length or number of somatic hypermutations (SHMs) as well as guidance regarding the choice of methodology is still lacking. In this work, we evaluated the ability of six available methods to reconstruct full-length BCRs using one simulated and three experimental SMART-seq datasets. In addition, we validated that the BCRs assembled in silico recognize their intended targets when expressed as monoclonal antibodies. We observed that methods such as BALDR, BASIC and BRACER showed the best overall performance across the tested datasets and conditions, whereas only BASIC demonstrated acceptable results on very short read libraries. Furthermore, the de novo assembly-based methods BRACER and BALDR were the most accurate in reconstructing BCRs harboring different degrees of SHMs in the variable domain, while TRUST4, MiXCR and BASIC were the fastest. Finally, we propose guidelines to select the best method based on the given data characteristics.

Introduction of the 4-(4-bromophenyl)benzenesulfonyl group to hydrazide analogs of Ilomastat leads to potent gelatinase B (MMP-9) inhibitors with improved selectivity.

Hydrazide derivatives of Ilomastat, carrying either aryl groups or distinct alkyl and arylsulfonyl moieties were synthesized and evaluated for their MMP inhibitory activity. Potent and selective MMP-9 inhibition (IC(50)=3 nM) was observed for compound 3m (arylsulfonyl group: 4-(4-Br-C6H4)-C6H4-SO(2)-). Interaction with the S2 enzyme subsite is mainly responsible for the inhibitory properties of this derivative as confirmed by molecular docking computation.

Structural Insights into the Forward and Reverse Enzymatic Reactions in Human Adenine Phosphoribosyltransferase.

Phosphoribosyltransferases catalyze the displacement of a PRPP α-1'-pyrophosphate to a nitrogen-containing nucleobase. How they control the balance of substrates/products binding and activities is poorly understood. Here, we investigated the human adenine phosphoribosyltransferase (hAPRT) that produces AMP in the purine salvage pathway. We show that a single oxygen atom from the Tyr105 side chain is responsible for selecting the active conformation of the 12 amino acid long catalytic loop. Using in vitro, cellular, and in crystallo approaches, we demonstrated that Tyr105 is key for the fine-tuning of the kinetic activity efficiencies of the forward and reverse reactions. Together, our results reveal an evolutionary pressure on the strictly conserved Tyr105 and on the dynamic motion of the flexible loop in phosphoribosyltransferases that is essential for purine biosynthesis in cells. These data also provide the framework for designing novel adenine derivatives that could modulate, through hAPRT, diseases-involved cellular pathways.

Selective blockade of the hydrolysis of the endocannabinoid 2-arachidonoylglycerol impairs learning and memory performance while producing antinociceptive activity in rodents.

Monoacylglycerol lipase (MAGL) represents a primary degradation enzyme of the endogenous cannabinoid (eCB), 2-arachidonoyglycerol (2-AG). This study reports a potent covalent MAGL inhibitor, SAR127303. The compound behaves as a selective and competitive inhibitor of mouse and human MAGL, which potently elevates hippocampal levels of 2-AG in mice. In vivo, SAR127303 produces antinociceptive effects in assays of inflammatory and visceral pain. In addition, the drug alters learning performance in several assays related to episodic, working and spatial memory. Moreover, long term potentiation (LTP) of CA1 synaptic transmission and acetylcholine release in the hippocampus, two hallmarks of memory function, are both decreased by SAR127303. Although inactive in acute seizure tests, repeated administration of SAR127303 delays the acquisition and decreases kindled seizures in mice, indicating that the drug slows down epileptogenesis, a finding deserving further investigation to evaluate the potential of MAGL inhibitors as antiepileptics. However, the observation that 2-AG hydrolysis blockade alters learning and memory performance, suggests that such drugs may have limited value as therapeutic agents.

A novel strategy for designing specific gelatinase A inhibitors: potential use to control tumor progression.

Matrix metalloproteases (MMPs) are zinc endopeptidases deeply implicated in tumor progression. MMP inhibitors are attractive potential anti-cancer agent. Unfortunately, until now, clinical trials remain disappointing, that could be the result of a lack of selectivity. We propose second generation selective MMPs, directed toward gelatinase A (MMP-2), based on a non-hydroxamate Zn-ligand grafted on the galardin (ilomastat) skeleton.

Improved gelatinase a selectivity by novel zinc binding groups containing galardin derivatives.

The synthesis of several analogues of galardin, a MMP inhibitor, are presented with their in vitro inhibitory activity against MMP-1 and MMP-2. These compounds contain a distinct Zinc Binding Group (ZBG). Those having a 2-acylated-heterocycle as well as a 2-arylamide function do not exhibit a good inhibition/selectivity against the enzymes tested. On the contrary, those that are based on a hydrazide scaffold present potent selectivity for MMP-2 versus MMP-1.