Biology subjects

de Souza, G. A.

Publications and source records attributed to de Souza, G. A..

9 recordsLinked to original sources

bioRxiv2026

From Data Curation to Risk Reporting: A Pipeline for Polygenic Risk Scores

Polygenic risk scores (PRS) have emerged as a powerful tool for quantifying genetic susceptibility to complex traits and diseases. However, their calculation and interpretation require standardized data curation, robust statistical methods, and clear reporting strategies. In this work, we present an integrated pipeline designed to address these challenges. The pipeline begins with the construction of a curated genotype/phenotype database derived from public repositories, ensuring that only phenotypes with appropriate metadata, statistical distributions, and ethical suitability are retained. The final dataset comprises 2,346 phenotypes covering 38,256,468 unique SNPs. These phenotypes serve as the final analytical units for PRS calculation, risk stratification, and individual-level interpretation. The generated reports integrate sample-level results, phenotype categorization, risk classification, study references, and variant tables, providing a structured and interpretable output for end users. Together, the curated database and reporting framework establish a comprehensive toolbox for PRS analysis, enhancing reproducibility, transparency, and usability in both research and clinical contexts.

bioinformatics↗

bioRxiv2026

GTX-GUT: A Standardized Metagenomic Workflow for Gut Microbiome Profiling and Clinical Associations

The human gut microbiome plays a central role in host physiology and disease, yet metagenomic analysis pipelines remain fragmented across sample preparation, taxonomic classification, and clinical interpretation stages, complicating reproducibility and translational use. Here we present GTX-GUT, a fully automated, containerized Snakemake pipeline for 16S rRNA gut microbiome profiling that integrates quality control, taxonomic classification (QIIME2/DADA2 against Greengenes 13.8), diversity and compositional metrics benchmarked against a curated healthy reference population, enterotype classification, a clinical association module spanning 11 disease categories, and automated natural-language report generation. We validated the pipeline using the ZymoBIOMICS mock community, showing that BBDuk preprocessing substantially reduced genus-level quantification error (Mean Absolute Error reduced from 7.34 to 1.58 percentage points; Pearson's $r$ improved from 0.576 to 0.833). Application to a human sample from a patient with type 2 Diabetes Mellitus recovered a dysbiotic signature consistent with the literature, including reduced Firmicutes abundance, elevated Bacteroidetes and Proteobacteria, and a predominance of clinical associations within metabolic and gastrointestinal categories. These results demonstrate that GTX-GUT provides a reproducible, end-to-end framework linking raw sequencing data to clinically interpretable output, with direct applicability to research and translational microbiome studies.

bioinformatics↗

bioRxiv2026

SlytheRINs: using graph parameters and residue interaction networks to analyze protein dynamics and structural ensembles

Establishing the relationship between protein structure and functional behavior remains a significant challenge. The recognition that proteins are inherently dynamic, with functions often dependent on conformational changes, is increasingly accepted. Among computational approaches for elucidating protein properties, Residue Interaction Network (RIN) analysis has emerged as a powerful tool. However, conventional RIN analysis is constrained by its reliance on single, static protein structures, which fail to capture the flexibility inherent in dynamic protein folding transitions. To address these limitations, SlytheRINs is introduced as an interactive tool designed for comparative analysis of protein conformations via RINs. SlytheRINs enables dynamic ensemble analysis by decomposing interaction network data across multiple conformations of a single protein and providing detailed residue-interaction mapping across conformational changes via graph parameters in comparative plots. Applying these principles, the conformational variations of the wild-type and a pathogenic variant (G188R) of the human Glucose-6-Phosphatase (G6PC1) catalytic subunit were compared to identify fluctuations in both chemical interactions and graph features associated with conformational changes induced by the residue modification. The analyses identified key shifts in dynamic residue interactions in the protein variant that compromise substrate binding and the catalytic site, thereby elucidating the impact on G6PC1s dynamic behavior and the resulting activity loss. AvailabilityThe source code for SlytheRINs is available on GitHub (https://github.com/evomol-lab/SlytheRINs), while the web tool and documentation are available at https://slytherins.streamlit.app.

bioinformatics↗

bioRxiv2026

Intrinsically disordered insert from SH2D2A rewires CD19 CAR signaling via Tyr290

Chimeric antigen receptor (CAR) T cells have transformed cancer immunotherapy, yet their truncated or suboptimal intracellular signaling can limit therapeutic efficacy. To enhance proximal signaling of a CD19-targeted CAR, we systematically inserted short Lck-recruiting motifs derived from Lck-adaptor proteins into the CAR intracellular tail. Six candidate sequences from four adaptor molecules (SH2D2A, SKAP1, LAT, LIME), with a sequence from CD3{varepsilon}, known to affect CAR functionality, as a positive control, were tested for expression and functional impact. Three CAR constructs (containing SH2D2A, LAT and LIME1 sequences respectively) displayed reduced surface expression, but only SH2D2A elicited a pronounced rewiring of CAR T cell phenotype following co-culture with CD19+ tumor lines. SH2D2A CAR T cells showed increased CD27 and CD56 expression and reduced expression of effector-associated mediators including granzyme B, IL-2, TNF, and IFN{gamma}. Through systematic mutagenesis and comparative phenotyping of SH2D2A CAR variants, we identified SH2D2A tyrosine 290 (Tyr290) as the critical residue mediating both the altered signaling phenotype and the low surface expression. Additionally, mutation of Tyr254 in the LIME1 CAR restored surface expression in Jurkat T cells, indicating insert- and context-dependent effects on receptor surface expression. Collectively, these results demonstrate that short, intrinsically disordered adaptor-derived sequences -- and single tyrosine residues within them -- can profoundly reprogram CAR signaling and expression.

immunology↗

bioRxiv2025

Beyond the Shared Inflammatory Axis: Differentiating Molecular Signatures in Psoriatic Arthritis and Ankylosing Spondylitis through Integrated Omics

Spondyloarthropathies (SpA), which include ankylosing spondylitis (AS) and psoriatic arthritis (PsA), are inflammatory rheumatic diseases with overlapping clinical and immunological characteristics but distinct molecular mechanisms. This study aimed to compare transcriptomic signatures of AS and PsA by investigating differentially expressed genes and associated regulatory pathways. Four public RNA-Seq datasets from GEO (GSE186063, GSE117769, GSE205748, and GSE221786) were used. Using an integrative strategy involving scientific text mining, functional enrichment, and biological curation, 433 genes with potential relevance to both conditions were selected. Based on this set, differential gene expression (DEG) analyses were performed, protein-protein interaction (PPI) networks and gene regulatory networks (GRN) were constructed, and hubs and master regulators (MRs) were identified. The integration of these approaches allowed the identification of shared inflammatory pathways, especially related to the IL-23/IL-17 axis, as well as disease-specific signatures. Functional analysis revealed enrichment in Th17 cell differentiation processes, lipopolysaccharide response, and activation of classical inflammatory pathways. This study provides a comprehensive comparative view of the transcriptomic signatures of PsA and AS that point to potential biomarkers and distinct therapeutic targets.

bioinformatics↗

bioRxiv2023

A novel approach to identify cross-identity peptides between Epstein-Barr virus and central nervous system proteins in Guillain-Barre syndrome and multiple sclerosis

BackgroundGuillain-Barre Syndrome (GBS) and multiple sclerosis are autoimmune diseases associated with an immune system attack response against peripheral and central nervous system autoantigens, respectively. Given the potential of Epstein-Barr virus (EBV) as a risk factor for both multiple sclerosis and GBS, the present study aimed to identify crucial residues among potential EBV CD4+ T lymphocyte epitopes and nervous system proteins. MethodsPublic databases (Allele Frequency Net Database, Immune Epitope Database, Genevestigator and Protein Atlas) were used to select proteins abundant in the nervous system, EBV immunogenic proteins, and HLA haplotypes. Computational tools were employed for predicting HLA-binding peptides and immunogenicity. For this, we developed immuno-cross, a Python tool (https://github.com/evoMOL-Lab/immuno-cross) to compare residue identity among nonamers. ResultsWe found ten proteins from the nervous system and 28 from EBV, which were used for predicting the binding peptides of 21 common HLAs in the world population. A total of 1411 haplotypes were distributed among 51 pairs of HLAs. Simulations were performed to determine whether nonamers from the EBV and nervous system proteins targeted TCR-contact residues. Then, three selection criteria were used, based on the relevance of each contact in the TCR-peptide-MHC interaction. The primary contact has to be located at position P5, and the positions P2, P3, and P8 were weighed as secondary, and P4, P6, and P7 were considered tertiary. Nonamers of EBV proteins and myelin proteins were combined in pairs and compared based on predefined selection criteria. The Periaxin protein had the highest number of nonamers pairs among PNS proteins, with 35 pairs. Four nonamers pairs from APLP1, two from CNP, and two from MBP bind to alleles of the haplotype DR-15. ConclusionsThe new approach proposed herein revealed that peptides derived from nervous system and EBV proteins share identical residues at critical contact points, which supports molecular mimicry. These findings suggest cross-reactivity between them and that the nonamer pairs identified with this approach have the potential to be an autoantigen. Experimental studies are needed to validate these findings.

bioinformatics↗

bioRxiv2019

EpitoCore: mining conserved epitope vaccine candidates in the core proteome of multiple bacteria strains

In reverse vaccinology approaches, complete proteomes of bacteria are submitted to multiple computational prediction steps in order to filter proteins that are possible vaccine candidates. Most available tools perform such analysis only in a single strain, or a very limited number of strains. But the vast amount of genomic data had shown that most bacteria contain pangenomes, i.e. their genomic information contains core, conserved genes, and random accessory genes specific to each strain. Therefore, it is of the utmost importance to define core proteins, and also core epitopes, in reverse vaccinology methods. EpitoCore is a decision-tree pipeline developed to fulfill that need. It provides surfaceome prediction of proteins from related strains, defines clusters of core proteins within those, calculate the immunogenicity of such clusters, predicts epitopes for a given set of MHC alleles defined by the user, and then reports if epitopes are located extracellularly and if they are conserved among the core homologues. Pipeline performance is illustrated by mining peptide vaccine candidates in Mycobacterium avium hominissuis strains. From a total proteome of approximately 4,800 proteins per strain, EpitoCore mined 103 highly immunogenic core homologues located at cell surface, many of those related to virulence and drug resistance. Conserved epitopes identified among these homologues allows the users to define sets of peptides with potential to immunize the largest coverage of tested HLA alleles using peptide-based vaccines. Therefore, EpitoCore is able to provide automated identification of conserved epitopes in bacterial pangenomic datasets.

bioinformatics↗

bioRxiv2018

Building comprehensive MS-friendly databases for proteomic analysis of bacterial species of unknown genetic background

In proteomics, peptide information within mass spectrometry data from a specific organism sample is routinely challenged against a protein sequence database that best represent such organism. However, if the species/strain in the sample is unknown or poorly genetically characterized, it becomes challenging to determine a database which can represent such sample. Building customized protein sequence databases merging multiple strains for a given species has become a strategy to overcome such restrictions. However, as more genetic information is publicly available and interesting genetic features such as the existence of pan- and core genes within a species are revealed, we questioned how efficient such merging strategies are to report relevant information. To test this assumption, we constructed databases containing conserved and unique sequences for ten different species. Features that are relevant for probabilistic-based protein identification by proteomics were then monitored. As expected, increase in database complexity correlates with pangenomic complexity. However, Mycobacterium tuberculosis and Bortedella pertusis generated very complex databases even having low pangenomic complexity or no pangenome at all. This suggests that discrepancies in gene annotation is higher than average between strains of those species. We further tested database performance by using mass spectrometry data from eight clinical strains from Mycobacterium tuberculosis, and from two published datasets from Staphylococcus aureus. We show that by using an approach where database size is controlled by removing repeated identical tryptic sequences across strains/species, computational time can be reduced drastically as database complexity increases.

bioinformatics↗

bioRxiv2017

The phosphatase PTPRG controls FGFR1 activity and influences sensitivity to FGFR kinase inhibitors

FGFR1 represents an important target for precision medicine and a detailed molecular understanding of the target is important in order to increase the efficacy of FGFR inhibitors. We have here applied proximity labelling of FGFR1 in an osteosarcoma cell line to identify determinants of FGFR1 activity. Many known FGFR interactors were identified (e.g. FRS2, PLC{gamma}, RSK2, SHC4, SRC), but the data also suggested novel determinants. A strong hit in our screen was the tyrosine phosphatase PTPRG. We show that PTPRG and FGFR1 interact and colocalize at the plasma membrane where PTPRG directly dephosphorylates activated FGFR1. We further show that osteosarcoma cell lines depleted for PTPRG display increased FGFR activity and are hypersensitive to stimulation by FGF1. In addition, PTPRG depletion elevated cell growth and negatively affected the efficacy of FGFR kinase inhibitors. Thus, PTPRG may have future clinical relevance by being a predictor of outcome after FGFR inhibitor treatment.

cell biology↗