Biology subjects

Righelli, D.

Publications and source records attributed to Righelli, D..

11 recordsLinked to original sources

bioRxiv2026

Interpreting Graph Neural Networks for Antibacterial Drug Repurposing against Acinetobacter baumannii

Antimicrobial resistance in Acinetobacter baumannii, whose carbapenem-resistant strains are classified by the World Health Organization as a critical priority, has motivated machine-learning screens for new antibacterials, but translating predictions into chemical hypotheses requires evaluating the molecular features on which a model relies. A graph neural network ensemble achieves competitive predictive performance against a descriptor-augmented reference model on the same antibacterial screen while using only atom features and connectivity, allowing interpretation to focus on the molecular graph without a separate input of handcrafted descriptors. The atom-level explanations identify features supporting the model's activity predictions under perturbation and restoration, while comparisons with equally sized perturbations within each molecule distinguish piperazine as the most consistently supported recurring contributor to predicted activity. Prospective screening of the Drug Repurposing Hub followed by developability and safety assessment yields a ranked shortlist for experimental follow-up and leads to testing of CE3F4, which reduced A. baumannii growth by 57% at 0.25 mM in a preliminary assay with a non-monotonic concentration--response profile.

bioinformatics↗

bioRxiv2026

LLMsFold: Integrating Large Language Models and Biophysical Simulations for De Novo Drug Design

The discovery of novel small molecules is challenging because of the vastness of chemical space and the complexity of protein-ligand interactions, leading to low success rates and time-consuming workflows. Here, we present LLMsFold, a computational framework that combines Large Language Models (LLMs) and biophysical foundation tools to design and validate new small molecules targeting pathogenic proteins. The pipeline starts by identifying viable binding pockets on a target protein through geometry-based pocket detection. A 70-billion-parameter transformer model from the LlaMA family then generates candidate molecules as SMILES strings under prompt constraints that enforce drug-likeness. Each molecule is evaluated by Boltz-2, a diffusion-based model for protein-ligand co-folding that predicts bound 3D structure and binding affinity. Promising candidates are iteratively optimized through a reinforcement learning loop that prioritizes high predicted affinity and synthetic accessibility. We demonstrate the approach on two challenging targets: ACVR1 (Activin A Receptor Type 1), implicated in fibrodysplasia ossificans progressiva (FOP), and CD19, a surface antigen expressed on most B-cell lymphoma and leukemia cells. Top candidates show strong in silico binding predictions and favorable drug-like profiles. All code and models are made available to support reproducibility and further development.

bioinformatics↗

bioRxiv2025

SpaceTrooper: a quality control framework for imaging-basedspatial omics data

Quality control (QC) is a critical step in the analysis of imaging-based single-cell spatial omics data, yet standardized metrics tailored to these technologies are still lacking. Most existing QC approaches are adapted from single-cell sequencing workflows and rely on fixed thresholds, limiting their ability to capture complex artifacts arising from image processing, tissue morphology, and platform-specific effects. Here, we present SpaceTrooper, a data-driven QC framework that computes an integrated per-cell quality score by combining expression-derived and morphological features. Without relying on fixed thresholds, SpaceTrooper systematically identifies low-quality cells caused by segmentation errors, signal loss, elevated background, spatial distortions, and tissue-derived artifacts. Across diverse tissues, technologies, and modalities, SpaceTrooper robustly detects technical failures and substantially improves clustering and cell-type coherence relative to conventional QC strategies.

bioinformatics↗

bioRxiv2025

Orchestrating Spatial Transcriptomics Analysis with Bioconductor

Spatial transcriptomics technologies provide spatially-resolved measurements of gene expression through assays that can either target selected genes or capture transcriptome-wide expression profiles. The complexity and variability of these technologies and their associated data necessitate multi-step workflows integrating diverse computational methods and software packages. We provide a freely accessible, open-source, continuously updated and tested online book containing reproducible code examples, datasets, and discussion about data analysis workflows for spatial omics data using Bioconductor in R, including interoperability with Python.

bioinformatics↗

bioRxiv2023

A Global Transcriptional Atlas of the Effect of Sleep Deprivation in the Mouse Frontal Cortex

Sleep deprivation (SD) has negative effects on brain function. Sleep problems are prevalent in neurodevelopmental, neurodegenerative and psychiatric disorders. Thus, understanding the molecular consequences of SD is of fundamental importance in neuroscience. In this study, we present the first simultaneous bulk and single-nuclear (sn)RNA sequencing characterization of the effects of SD in the mouse frontal cortex. We show that SD predominantly affects glutamatergic neurons, specifically in layers 4 and 5, and produces isoform switching of thousands of transcripts. At both the global and cell-type specific level, SD has a large repressive effect on transcription, down-regulating thousands of genes and transcripts; underscoring the importance of accounting for the effects of sleep loss in transcriptome studies of brain function. As a resource we provide extensive characterizations of cell types, genes, transcripts and pathways affected by SD; as well as tutorials for data analysis.

neuroscience↗

bioRxiv2021

Curated Single Cell Multimodal Landmark Datasets for R/Bioconductor

BackgroundThe majority of high-throughput single-cell molecular profiling methods quantify RNA expression; however, recent multimodal profiling methods add simultaneous measurement of genomic, proteomic, epigenetic, and/or spatial information on the same cells. The development of new statistical and computational methods in Bioconductor for such data will be facilitated by easy availability of landmark datasets using standard data classes. ResultsWe collected, processed, and packaged publicly available landmark datasets from important single-cell multimodal protocols, including CITE-Seq, ECCITE-Seq, SCoPE2, scNMT, 10X Multiome, seqFISH, and G&T. We integrate data modalities via the MultiAssayExperiment Bioconductor class, document and re-distribute datasets as the SingleCellMultiModal package in Bioconductors Cloud-based ExperimentHub. The result is single-command actualization of landmark datasets from seven single-cell multimodal data generation technologies, without need for further data processing or wrangling in order to analyze and develop methods within Bioconductors ecosystem of hundreds of packages for single-cell and multimodal data. ConclusionsWe provide two examples of integrative analyses that are greatly simplified by SingleCellMultiModal. The package will facilitate development of bioinformatic and statistical methods in Bioconductor to meet the challenges of integrating molecular layers and analyzing phenotypic outputs including cell differentiation, activity, and disease. Author SummaryExperimental data packages that provide landmark datasets have historically played an important role in the development of new statistical methods in Bioconductor by lowering the barrier of access to relevant data, providing a common testing ground for software development and benchmarking, and encouraging interoperability around common data structures. In this manuscript, we review major classes of technologies for collecting multimodal data including genomics, transcriptomics, epigenetics, proteomics, and spatial information at the level of single cells. We present the SingleCellMultiModal R/Bioconductor package that provides single-command access to landmark datasets from seven different technologies, storing datasets using HDF5 and sparse arrays for memory efficiency and integrating data modalities via the MultiAssayExperiment class. We demonstrate two integrative analyses that are greatly simplified by SingleCellMultiModal. The package facilitates development and benchmarking of bioinformatic and statistical methods to integrate molecular layers at the level of single cells with phenotypic outputs including cell differentiation, activity, and disease, within Bioconductors ecosystem of hundreds of packages for single-cell and multimodal data.

bioinformatics↗

bioRxiv2021

SpatialExperiment: infrastructure for spatially resolved transcriptomics data in R using Bioconductor

SummarySpatialExperiment is a new data infrastructure for storing and accessing spatially resolved transcriptomics data, implemented within the R/Bioconductor framework, which provides advantages of modularity, interoperability, standardized operations, and comprehensive documentation. Here, we demonstrate the structure and user interface with examples from the 10x Genomics Visium and seqFISH platforms, and provide access to example datasets and visualization tools in the STexampleData, TENxVisiumData, and ggspavis packages. Availability and ImplementationThe SpatialExperiment, STexampleData, TENxVisiumData, and ggspavis packages are available from Bioconductor. The package versions described in this manuscript are available in Bioconductor version 3.15 onwards. Contactrisso.davide@gmail.com, shicks19@jhu.edu Supplementary InformationSupplementary Tables and Figures are available online.

bioinformatics↗

bioRxiv2020

Single cell multi-omic analysis identifies a Tbx1-dependent multilineage primed population in the murine cardiopharyngeal mesoderm

The poles of the heart and branchiomeric muscles of the face and neck are formed from the cardiopharyngeal mesoderm (CPM) within the pharyngeal apparatus. The formation of the cardiac outflow tract and branchiomeric muscles are disrupted in patients with 22q11.2 deletion syndrome (22q11.2DS), due to haploinsufficiency of TBX1, encoding a T-box transcription factor. Here, using single cell RNA-sequencing, we identified a multilineage primed population (MLP) within the CPM, marked by the Tbx1 lineage, which has bipotent properties to form cardiac and skeletal muscle cells. The MLPs are localized within the nascent mesoderm of the caudal lateral pharyngeal apparatus and provide a continuous source of progenitors that undergo TBX1-dependent progression towards maturation. Tbx1 also regulates the balance between MLP maintenance and maturation while restricting ectopic non-mesodermal gene expression. We further show that TBX1 confers this balance by direct regulation of MLP enriched genes and downstream pathways, partly through altering chromatin accessibility. Our study thus uncovers a new cell population and reveals novel mechanisms by which Tbx1 directs the development of the pharyngeal apparatus, which is profoundly altered in 22q11.2DS.

developmental biology↗

bioRxiv2020

Easyreporting simplifies the implementation of Reproducible Research Layers in R software

During last years "irreproducibility" became a general problem in omics data analysis due to the use of sophisticated and poorly described computational procedures. For avoiding misleading results, it is necessary to inspect and reproduce the entire data analysis as a unified product. Reproducible Research (RR) provides general guidelines for public access to the analytic data and related analysis code combined with natural language documentation, allowing third-parties to reproduce the findings. We developed easyreporting, a novel R/Bioconductor package, to facilitate the implementation of an RR layer inside reports/tools without requiring any knowledge of the R Markdown language. We describe the main functionalities and illustrate how to create an analysis report using a typical case study concerning the analysis of RNA-seq data. Then, we also show how to trace R functions automatically. Thanks to this latter feature, easyreporting results beneficial for developers to implement procedures that automatically keep track of the analysis steps within Graphical User Interfaces (GUIs). Easyreporting can be useful in supporting the reproducibility of any data analysis project and the implementation of GUIs. It turns out to be very helpful in bioinformatics, where the complexity of the analyses makes it extremely difficult to trace all the steps and parameters used in the study.

bioinformatics↗

bioRxiv2020

Chromatin and transcriptional response to loss of TBX1 in differentiating mouse P19Cl6 and embryonic stem cells.

The T-box transcription factor TBX1 has critical roles in the cardiopharyngeal lineage and the gene is haploinsufficient in DiGeorge syndrome, a typical developmental anomaly of the pharyngeal apparatus. Despite almost two decades of research, if and how TBX1 function triggers chromatin remodeling is not known. Here, we explored genome-wide gene expression and chromatin remodeling in two independent cellular models of Tbx1 loss of function, mouse embryonic carcinoma cells P19Cl6, and mouse embryonic stem cells (mESCs). The results of our study revealed that the loss or knockdown of TBX1 caused extensive transcriptional changes, some of which were cell type-specific, some were in common between the two models. However, unexpectedly we observed only limited chromatin changes in both systems. In P19Cl6 cells, differentially accessible regions (DARs) were not enriched in T-BOX binding motifs; in contrast, in mESCs, 34% (n=47) of all DARs included a T-BOX binding motif and almost all of them gained accessibility in Tbx1-/- cells. In conclusion, despite a clear transcriptional response of our cell models to loss of TBX1 in early cell differentiation, chromatin changes were relatively modest.

genomics↗

bioRxiv2018

Shank3 Modulates Sleep and Expression of Circadian Transcription Factors

Autism Spectrum Disorder (ASD) is the most prevalent neurodevelopmental disorder in the United States and often co-presents with sleep problems. Sleep problems in ASD predict the severity of ASD core diagnostic symptoms and have a considerable impact on the quality of life of caregivers. Little is known, however, about the underlying molecular mechanisms. We investigated the role of Shank3, a high confidence ASD gene candidate, in sleep architecture and regulation. We show that mice lacking exon 21 of Shank3 have problems falling asleep even when sleepy. Using RNA-seq we show that sleep deprivation increases the differences in gene expression between mutants and wild types, downregulating circadian transcription factors Per3, Dec2, Hlf, Tef, and Reverb. Shank3 mutants also have trouble regulating wheel-running activity in constant darkness. Overall our study shows that Shank3 is an important modulator of sleep and clock gene expression.

neuroscience↗