Biology subjects

Taccioli, C.

Publications and source records attributed to Taccioli, C..

10 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

Five hundred million years of methylation: tracing the mutational origins of vertebrate genome composition

Methylation-associated deamination removes CpG from vertebrate genomes, but how it affects the dinucleotide profile remains unresolved. We analysed 753 vertebrate and 481 invertebrate genomes to test whether CpG loss defines a compositional axis and to identify its strongest signature. CpG depletion was the dominant axis of vertebrate dinucleotide variation. Unexpectedly, its strongest between-genome correlate was AG/CT rather than the direct mutational product TpG/CpA, which showed the expected dataset-wide mass balance but varied little among genomes. A forward-evolution model based on measured seven-nucleotide human germline substitution rates produced neither CpG depletion nor AG/CT enrichment when methylated-CpG mutability was excluded. Adding one CpG-specific mutability term, calibrated only to the mammalian CpG ratio, reproduced both features, identifying AG/CT as a second-order consequence of the context-dependent mutation network. Within genomes, CpG depletion was strongest in transposable elements and weakened with distance from them. Across vertebrates, the axis followed Amniota more closely than endothermy and was associated with an expanded GC-rich isochore compartment. A Machine Learning analysis shows that CpG depletion and AG/CT were the principal features separating vertebrates from invertebrates, in which both were markedly attenuated. Thus, a methylation-associated axis organises vertebrate dinucleotide composition, and its strongest marker is not the immediate product of CpG deamination.

genomics↗

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↗

bioRxiv2024

GBRAP: A Comprehensive Database and Tool for Exploring Genomic Diversity Across All Domains of Life

Evolutionary studies require extensive examination of genomic information across all domains of life. Despite the availability of a large number of genomes through GenBank, the effective visualisation or comparison of the information they contain is challenging due to many reasons including their size. We introduce GBRAP, a comprehensive software tool to analyse GenBank files, and an online database housing an extensive collection of carefully curated, high-quality genome statistics for all the organisms available in the RefSeq database of NCBI. Users can either directly search, or select from pre-categorized groups, the organisms of their choice and retrieve data, which eventually output as tables containing more than 200 columns of useful genomic information (Base counts, GC content, Shannon Entropy, Codon Usage etc.) separately calculated for different genomic elements (eg. CDS, Introns, tRNA, rRNA, ncRNA, etc.). The data are independently displayed (if applicable) for each chromosomal, mitochondrial, plastid, or plasmid sequence. All the data can be visualised on the database or downloaded as CSV or Excel files. The GBRAP database is free to access without any registration and is publicly available at .

genomics↗

bioRxiv2024

TumFlow: An AI Model for Predicting New Anticancer Molecules

MotivationMelanoma is a severe form of skin cancer increasing globally with about 324.000 cases in 2020, making it the fifth most common cancer in the United States. Conventional drug discovery methods face limitations due to the inherently time consuming and costly. However, the emergence of artificial intelligence (AI) has opened up new possibilities. AI models can effectively simulate and evaluate the properties of a vast number of potential drug candidates, substantially reducing the time and resources required by traditional drug discovery processes. In this context, the development of AI normalizing flow models, employing machine learning techniques to create new molecular structures, holds great promise for accelerating the discovery of effective anticancer therapies. ResultsThis manuscript introduces a novel AI model, named TumFlow, aimed at generating new molecular entities with potential therapeutic value in cancer treatment. It has been trained on the comprehensive NCI-60 dataset, encompassing thousands of molecules tested across 60 tumour cell lines, with a specific emphasis on the melanoma SK-MEL-28 cell line. The model successfully generated new molecules with predicted improved efficacy in inhibiting tumour growth while being synthetically feasible. This represents a significant advancement over conventional generative models, which often produce molecules that are challenging or impossible to synthesize. Furthermore, TumFlow has also been utilized to optimize molecules known for their efficacy in clinical melanoma treatments. This led to the creation of novel molecules with a predicted enhanced likelihood of effectiveness against melanoma, currently undocumented on PubChem. Availability and Implementationhttps://github.com/drigoni/TumFlow. Supplementary informationUploaded.

bioinformatics↗

bioRxiv2021

Involvement of non-LTR retrotransposons in the cancer incidence and lifespan of mammals

The presence in nature of closely related species showing drastic differences in lifespan and cancer incidence has recently increased the interest of the scientific community on these topics. In particular, the adaptations and genomic characteristics underlying the evolution of cancer-resistant and long-lived species have recently focused on the presence of alterations in the number of non-coding RNAs, on epigenetic regulation and, finally, on the activity of transposable elements (TEs). In this study, we compared the content and dynamics of TE activity in the genomes of four rodent and six bat species exhibiting different lifespans and cancer susceptibility. Mouse, rat and guinea pig (short-lived and cancer-prone organisms) were compared with the naked mole rat (Heterocephalus glaber) which is the rodent with the longest lifespan. The long-lived and cancer-resistant bats of the genera Myotis, Rhinolophus, Pteropus and Rousettus were instead compared with the Molossus, which is instead a short-lived and cancer-resistant organism. Analyzing the patterns of recent accumulations of TEs in the genome in these species, we found a strong suppression or negative selection to accumulation, of non-LTR retrotransposons in long-lived and cancer-resistant organisms. On the other hand, all short-lived and cancer-prone species have shown recent accumulation of this class of TEs. Among bats, the Molossus molossus turned out to be a very particular species and, at the same time, an important model because, despite being susceptible to rapid ageing, it is resistant to cancer. In particular, we found that its genome has the highest density of SINE (non-LTR retrotransposons), but, on the other hand, a total lack of active LINE retrotransposons. Our hypothesis is that the lack of LINEs presumably makes the Molossus cancer resistant due to lack of retrotransposition but, at the same time, the high presence of SINE, may be related to their short life span due to "sterile inflammation" and high mutation load. We suggest that research on ageing and cancer evolution should put particular attention to the involvement of non-LTR retrotransposons in these phenomena.

evolutionary biology↗

bioRxiv2021

miRNAs Copy Number Variations repertoire as hallmark indicator of cancer species predisposition

Aging is one of the hallmarks of multiple human diseases, including cancer. However, the molecular mechanisms associated with high longevity and low cancer incidence percentages characterizing long-living organisms have not been fully understood yet. In this context, we hypothesized that variations in the number of copies (CNVs) of specific genes may protect some species from cancer onset. Based on the statistical comparison of gene copy numbers within the genomes of cancer -prone and -resistant organisms, we identified novel gene targets linked to the tumor predisposition of a species, such as CD52, SAT1 and SUMO protein family members. Furthermore, for the first time, we were able to discover that, considering the entire genome copy number landscape of a species, microRNAs (miRNAs) are among the most significant gene families enriched for cancer progression and predisposition. However, their roles in ageing and cancer resistance from a comparative perspective remains largely unknown. To this end, we identified through bioinformatics analysis, several alterations in miRNAs copy number patterns, represented by duplication of miR-221, miR-222, miR-21, miR-372, miR-30b, miR-30d and miR-31 among others. Therefore, our analysis provides the first evidence that an altered copy number miRNAs signature is able to statistically discriminate species more susceptible to cancer than those that are tumor resistant, helping researchers to discover new possible therapeutic targets involved in tumor predisposition.

genomics↗

bioRxiv2021

GBRAP: a tool to retrieve, parse and analyze GenBank files of viral and bacterial species

SummaryGenBank files contain genomic data of sequenced living organisms. Here, we present GBRAP (GenBank Retrieving, Analyzing and Parsing software), a tool written in Python 3 that can be used to easily download, parse and analyze viral and bacterial GenBank files, even when contain more than one genomic sequence for each species. GBRAP can analyze more files simultaneously through single command-line parameters that give as output a single table showing the genomic characteristics of each organism. It is also able to calculate Shannon, LZSS (Lempel-Ziv-Storer-Szymanski) and topological entropy for both the entire genome and its constitutive elements such as genes, rRNAs, tRNAs, tmRNAs and ncRNAs together with Chargaffs second parity rule scores obtained using different mathematical methods. Moreover, GBRAP can calculate, the number, the length and the nucleotides abundance of genomic components for each DNA strand and for the overlapping regions among the two complementary helixes. To our knowledge, this is the only software capable of providing this type of genomic analyses all together in a single tool, that, therefore can be used by the scientists interested in both genomics and evolutionary research. Availability and implementationThe data underlying this article are available from the corresponding author on reasonable request.

bioinformatics↗

bioRxiv2019

DNA energy constraints shape biological evolutionary trajectories

Most living systems rely on double-stranded DNA (dsDNA) to store their genetic information and perpetrate themselves. Thus, the biological information contained within a dsDNA molecule, in terms of a linear sequence of nucleotides, has been considered the main target of the evolution. However, in this information-centred perspective, certain DNA sequence symmetries are difficult to explain. Here we show that these patterns can emerge from the physical peculiarities of the dsDNA molecule itself and the maximum entropy principle alone, rather than from biological or environmental evolutionary pressure. Our predictions are valid for both prokaryotes and eukaryotes, and also inform the interpretation of observed codon biases and context-dependent mutation patterns in human populations. Our results suggest that the double helix energy constraints and, more generally, the physical properties of the dsDNA are the hard drivers of the overall DNA sequence architecture, whereas the biological selective processes act as soft drivers, which only under extraordinary circumstances overtake the overall entropy content of the genome.

molecular biology↗

bioRxiv2017

Impact of non-LTR retrotransposons in the differentiation and evolution of Anatomically Modern Humans

Transposable Elements are biologically important components of eukaryote genomes. In particular, non-LTR retrotransposons (N-LTRrs) extensively shaped the human genome throughout evolution. In this study, we compared retrotransposon insertions differentially present in the genomes of Anatomically Modern Humans, Neanderthals, Denisovans and Chimpanzees, in order to assess the possible impact of retrotransposition in the differentiation of the human lineage. Briefly, we first identified species-specific N-LTRrs and established their distribution in present day human populations. These analyses shortlisted a group of N-LTRr insertions that were found exclusively in Anatomically Modern Humans. Notably, these insertions targeted genes more frequently than randomly expected and are associated with an increase in the number of transcriptional/splicing variants of those genes they inserted in. The analysis of the functionality of genes targeted by human-specific N-LTRr insertions seems to reflect phenotypic changes that occurred during human evolution. Furthermore, the expression of genes containing the most recent N-LTRr insertions is enriched in the brain, especially in undifferentiated neurons, and these genes associate in networks related to neuron maturation and migration. Additionally, we also identified candidate N-LTRr insertions that have likely produced new functional variants exclusive to modern humans, which show traces of positive selection and are now fixed in all present-day human populations. In sum, our results strongly suggest that N-LTRr impacted our differentiation as a species and have been a constant source of genomic variability all throughout the evolution of the human lineage.

genomics↗