Biology subjects

Riccardi, C.

Publications and source records attributed to Riccardi, C..

9 recordsLinked to original sources

bioRxiv2025

Supervised white matter bundle segmentation in glioma patients with transfer learning

In clinical settings, the virtual dissection of white matter tracts represents an informative source of information for monitoring neurological conditions or to support the planning of a treatment. The automation of this task using data-driven approaches, and in particular deep learning models, provides promising evidence of good accuracy when carried out on healthy individuals. However, the lack of large clinical datasets and the profound differences between healthy and clinical populations hinder the translation of these results to patients. Here, we investigated for the first time the effectiveness of transfer learning in adapting a deep learning architecture trained on a healthy population to glioma patients. Importantly, we provided the first thorough characterization of domain shift and its complexity, distinguishing systematic (i.e. measurement and pre-processing related) from tumor specific components. Our results suggest that (i) models trained on a large normative healthy population have a significant performance drop when the inference is carried out on patients; (ii) transfer learning can be an effective strategy to overcome the shortage of clinical data and to manage the systematic shift; (iii) fine tuning of the learning model cannot accommodate large white matter deformations induced by the tumor. The results were coherent across the five white matter bundles and three different input modalities tested, highlighting their robustness and generalizability. Our work provides valuable insights for advancing automated white matter segmentation in clinical populations and enhancing clinical transfer learning applications. Graphical Abstract O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=72 SRC="FIGDIR/small/659452v1_ufig1.gif" ALT="Figure 1"> View larger version (16K): org.highwire.dtl.DTLVardef@1c2bbbcorg.highwire.dtl.DTLVardef@1b3555eorg.highwire.dtl.DTLVardef@a9bb4dorg.highwire.dtl.DTLVardef@147dc5e_HPS_FORMAT_FIGEXP M_FIG C_FIG HighlightsO_LIWhite matter bundle segmentation is essential for planning glioma tumor resection. C_LIO_LIWe trained a deep learning model on a healthy dataset to segment white matter bundles. C_LIO_LIWe observed a sub-optimal performance on glioma patients due to domain shift. C_LIO_LITransfer learning mitigated domain shift between healthy and clinical populations. C_LIO_LIWe highlighted domain shift domains solved/not solved by transfer learning. C_LI

neuroscience↗

bioRxiv2024

Engineering the Marine Pseudoalteromonas haloplanktis TAC125 via pMEGA Plasmid Targeted Curing Using PTasRNA Technology

Marine bacteria that have adapted to thrive in extreme environments, such as Pseudoalteromonas haloplanktis TAC125 (PhTAC125), offer a unique biotechnological potential. The discovery of an endogenous megaplasmid (pMEGA) raised questions about its metabolic impact and functional role in this strain. This study aimed at streamlining the host genetic background by curing PhTAC125 from the pMEGA plasmid using a sequential genetic approach. We combined homologous recombination by exploiting a suicide vector with the PTasRNA gene silencing technology to interfere with pMEGA replication machinery. This approach led to the construction of the novel PhTAC125 KrPL2 strain, cured from the pMEGA plasmid, which exhibited no significant differences in the growth behaviour, though showcasing enhanced resistance to oxidative stress and a reduced capability of biofilm formation. These findings represent a significant achievement for understanding of the role of pMEGA plasmid and for the biotechnological applications of PhTAC125 in recombinant protein production. This opens up the possibility to exploit pMEGA valuable genetic elements and further advancing the genetic tools for PhTAC125.

microbiology↗

bioRxiv2024

The acquisition of additional control over quorum sensing regulation buffers noise in microbial growth dynamics

Quorum sensing (QS) is a cell-to-cell communication system used by bacteria to act collectively. Often, bacteria possess more than one QS regulatory module that form complex regulatory networks. Presumably, these configurations have evolved through the integration of novel transcription factors into the native regulatory systems. The selective advantages provided by these alternative configurations on QS-related phenotypes is poorly predictable only based on their underlying network structure. Here we show that the acquisition of extra regulatory modules of QS has important consequences on the overall regulation of microbial growth dynamics by significantly reducing the variability in the final size of the population in Burkholderia. We mapped the distribution of horizontally transferred QS modules in extant bacterial genomes, finding that these tend to add up to already-present modules in the majority of cases, 63.32%. We then selected a strain harboring two intertwined QS modules and,using mathematical modelling, we predicted an intrinsic ability of the newly acquired module to buffer noise in growth dynamics. We experimentally validated this prediction choosing one strain possessing both systems, deleting one of the two and measuring key growth parameters and QS synthase expression. We extended such considerations on two other strains naturally implementing the two versions of the QS regulation studied herein. Finally, using transcriptomics, we show that the de-regulation of metabolism likely plays a key role in differentiating the two configurations. Our results shed light on the role of additional control over QS regulation and illuminate on the possible phenotypes that may arise after HGT events.

systems biology↗

bioRxiv2024

Reduction of fungal dysbiosis is involved in the attenuation of Dextran Sodium Sulfate-Induced Mouse Colitis Mediated by GILZ protein and yeast extract compound

Inflammatory bowel diseases (IBD), such as Crohns disease and ulcerative colitis, have a complex and multifactorial pathogenesis that remains not fully elucidated. Recent research suggests that intestinal fungal dysbiosis may contribute to the development and persistence of IBD. In this study, we explored, for the first time, the effects of the glucocorticoid-induced leucine zipper (GILZ) protein, known to have protective effects on the gut mucosa in preclinical IBD models, in combination with a yeast extract, which supports the growth of beneficial microorganisms, in a mouse model of ulcerative colitis. The combined treatment produced significant protection against severe disease outcomes in the mice, including the restoration of intestinal barrier integrity and the reduction of pro-inflammatory cytokines. Specifically, GILZ primarily acted on the gut permeability, while the yeast extract mainly reduced pro-inflammatory cytokines. Notably, both treatments were effective in restoring the intestinal burden of clinically important Candida and former Candida species. Analysis of the intestinal fungal communities revealed that both treatments were able to reduce colitis-associated fungal dysbiosis, promoting a fungal composition similar to that of healthy mice. This effect was mainly the result of a decreased abundance of the Meyerozima genus, which was dominant in the colitic mice. Thus, combined treatment regimens with the GILZ protein and yeast extract could represent a new strategy for the treatment of inflammatory bowel diseases, by targeting multiple mechanisms at the basis of IBD, including the fungal dysbiosis. IMPORTANCEInflammatory bowel diseases (IBD), including Crohns disease and ulcerative colitis, are characterized by chronic inflammation and have a complex, multifactorial pathogenesis that is not yet fully understood. Currently, no established therapeutic strategy can consistently manage IBD effectively. Recent research indicates that intestinal fungal dysbiosis could potentially contribute to the development and persistence of chronic IBD, highlighting the importance of investigating alternative therapeutic strategies able to attenuate fungal dysbiosis in the context of intestinal inflammation. In this study, we demonstrate that a combination of a recombinant protein (GILZp) and a compound with prebiotic properties could represent a new therapeutic strategy for the treatment of IBD, as it not only decreases inflammation and restores the integrity of the epithelial barrier, but reduces fungal dysbiosis associated with DSS-induced colitis.

microbiology↗

bioRxiv2023

Supervised White Matter Bundle Segmentation in Glioma Patients with Transfer Learning

In clinical neuroscience, the segmentation of the main white matter bundles is propaedeutic for many tasks such as preoperative neurosurgical planning and monitoring of neuro-related diseases. Automating bundle segmentation with data-driven approaches and deep learning models has shown promising accuracy in the context of healthy individuals. The lack of large clinical datasets is preventing the translation of these results to patients. Inference on patients data with models trained on healthy population is not effective because of domain shift. This study aims to carry out an empirical analysis to investigate how transfer learning might be beneficial to overcome these limitations. For our analysis, we consider a public dataset with hundreds of individuals and a clinical dataset of glioma patients. We focus our preliminary investigation on the corticospinal tract. The results show that transfer learning might be effective in partially overcoming the domain shift.

neuroscience↗

bioRxiv2022

Metabolic robustness to growth temperature of cold adapted bacterium

Microbial communities experience continuous environmental changes, among which temperature fluctuations are arguably the most impacting. This is particularly important considering the ongoing global warming but also in the "simpler" context of seasonal variability of sea-surface temperature. Understanding how microorganisms react at the cellular level can improve our understanding of possible adaptations of microbial communities to a changing environment. In this work, we investigated which are the mechanisms through which metabolic homeostasis is maintained in a cold-adapted bacterium during growth at temperatures that differ widely (15 and 0{degrees}C). We have quantified its intracellular and extracellular central metabolomes together with changes occurring at the transcriptomic level in the same growth conditions. This information was then used to contextualize a genome-scale metabolic reconstruction and to provide a systemic understanding of cellular adaptation to growth at two different temperatures. Our findings indicate a strong metabolic robustness at the level of the main central metabolites, counteracted by a relatively deep transcriptomic reprogramming that includes changes in gene expression of hundreds of metabolic genes. We interpret this as a transcriptomic buffering of cellular metabolism, able to produce overlapping metabolic phenotypes despite the wide temperature gap. Moreover, we show that metabolic adaptation seems to be mostly played at the level of few key intermediates (e.g. phosphoenolpyruvate) and in the cross-talk between the main central metabolic pathways. Overall, our findings reveal a complex interplay at gene expression level that contributes to the robustness/resilience of core metabolism, also promoting the leveraging of state-of-the-art multi-disciplinary approaches to fully comprehend molecular adaptations to environmental fluctuations.

systems biology↗

bioRxiv2022

The selective force driving metabolic operon assembly

The evolution of operons has puzzled evolutionary biologists since their discovery and many theories exist to explain their emergence and spreading. The presence of several plausible hypotheses dealing with operon emergence/evolution/spreading is indicative of the absence of a universal causal factor for this evolutionary process. Here, we argue that the way in which DNA replication and cell division are coupled in microbial species introduces an additional selective force that may be responsible for the clustering of functionally related genes on chromosomes. We interpret this as a preliminary and necessary step in operon formation. Specifically, we start from the observation that during DNA replication differences in copy number of genes that are found at distant loci on the same chromosome arm exist. We provide theoretical considerations suggesting that, when genes of the same metabolic process are far away on the chromosome, this results in perturbations to metabolic homeostasis. By formalizing the effect of DNA replication on metabolic homeostasis based on Metabolic Control Analysis, we show that the above situation provides a selective force that can drive the formation of gene clusters and operons. Finally, we confirmed that, in present-day genomes, this force is significantly stronger in those species where the average number of active replication forks is larger and quantify the theoretical contribution of this feature on the distribution of extant gene clusters and operons.

systems biology↗

bioRxiv2022

Modelling the metabolic consequences of antimicrobial exposure

Besides genetic mutations, the metabolic state of bacterial cells represents another driving factor in the emergence of antimicrobial resistance and in the actual efficacy of treatments. In this direction, studying how bacteria reprogram their metabolism when facing antimicrobial exposure is crucial to enhance our ability to limit the development and spread of antibiotic resistance. Here we have studied the metabolic consequences of antimicrobial exposure in bacteria using an integrated approach that exploits transcriptomics and computational modelling. Specifically, we asked whether common metabolic strategies emerge during the exposure to antimicrobials, regardless of the kind of antimicrobial used or, on the contrary, antimicrobial-specific pathways exist. To this purpose, we have used an heterogeneous dataset from six published studies on Escherichia coli exposed to different concentrations/types of compounds. We show that experimental condition, not antimicrobial exposure, is the factor that influences the most the resulting metabolic networks. However, despite condition-dependent metabolic signatures being evident, specific changes in flux distributions by antimicrobial exposed cells could be identified. In particular, purine and pyrimidine biosynthesis, and cofactor and prosthetic group biosynthesis were commonly affected by all considered antimicrobials. This suggests the presence of general metabolic strategies to face the stress posed by antimicrobial exposure and that, in turn, may represent an untapped resource for the fight against microbial infections. Finally, our analysis predicted an overall metabolic rewiring following bacteriostatic vs. bactericidal drug exposure that is in line with the current knowledge about the effects of these two classes of compounds on microbial metabolic phenotypes. IMPORTANCEA mechanistic understanding of microbial metabolic reprogramming during antimicrobial exposure is key to facilitate the discovery of new resistance mechanisms and to identify novel areas of intervention to face microbial infections. This study shows how the integration of transcriptomic data and genome-scale metabolic modelling can be used to address this critical issue, and to trace general metabolic strategies exploited by bacteria to face the stress posed by antimicrobial drugs.

systems biology↗

bioRxiv2022

Extraction and analysis of methylation features from Pacific Biosciences SMRT reads using MeStudio

MotivationDNA methylation is the most relevant epigenetic information, present in eukaryotes and prokaryotes, and is related to several biological phenomena, from cellular differentiation to control of gene flow, pathogenesis and virulence. The widespread use of third-generation sequencing technologies allows direct and easy detection of genome-wide methylation profiles, offering increasing opportunities to understand and exploit the epigenomics landscape. ResultsWe introduce MeStudio, a pipeline which allows to analyse and combine genome-wide methylation profiles with genomic features. Outputs report the presence of DNA methylation in coding sequences, noncoding sequences, intergenic sequences, and sequences upstream to CDS. We show the usage and performances of MeStudio on a set of single-molecule real time sequencing outputs from the bacterial species Sinorhizobium meliloti. Availability and ImplementationMeStudio is written in Python, Bash and C and is freely available under an open source GPLv3 license at https://github.com/combogenomics/MeStudio Supplementary informationSupplementary data are available at Bioinformatics online. Contactcombo.unifi@gmail.com

bioinformatics↗