Biology subjects

Duarte, J.

Publications and source records attributed to Duarte, J..

7 recordsLinked to original sources

bioRxiv2025

Multi-scale structural similarity embedding search across entire proteomes

The rapid expansion of three-dimensional (3D) biomolecular structure information, driven by breakthroughs in artificial intelligence/deep learning (AI/DL)-based structure predictions, has created an urgent need for scalable and efficient structure similarity search methods. Traditional alignment-based approaches, such as structural superposition tools, are computationally expensive and challenging to scale with the vast number of available macromolecular structures. Herein, we present a scalable structure similarity search strategy designed to navigate extensive repositories of experimentally determined structures and computed structure models predicted using AI/DL methods. Our approach leverages protein language models and a deep neural network architecture to transform 3D structures into fixed-length vectors, enabling efficient large-scale comparisons. Although trained to predict TM-scores between single-domain structures, our model generalizes beyond the domain level, accurately identifying 3D similarity for full-length polypeptide chains and multimeric assemblies. By integrating vector databases, our method facilitates efficient large-scale structure retrieval, addressing the growing challenges posed by the expanding volume of 3D biostructure information.

bioinformatics↗

bioRxiv2024

Plant G x Microbial E: Plant genotype interaction with soil bacterial community shapes rhizosphere composition during invasion

It is increasingly recognized that different genetic variants can uniquely shape their microbiomes. Invasive species often evolve in their introduced ranges, but little is known about the potential for microbial associations to evolve during invasion as a result. We investigated invader genotype (G) and microbial environment (E) interactions in C. solstitialis (yellow starthistle), a Eurasian plant that is known to have evolved novel genotypes, and to have altered microbial interactions, in its severe invasion of California, USA. We conducted an experiment in which native and invading genotypes were inoculated with native and invaded range soil microbial communities. We used amplicon sequencing to characterize rhizosphere bacteria in both the experiment and the field soils from which they were derived. We found that bacterial diversity is higher in invaded soils, but that invading genotypes accumulated a lower diversity of bacteria and unique microbial composition in experimental inoculations, relative to native genotypes. Associations with potentially beneficial Streptomycetaceae were particularly interesting, as these were more abundant in the invaded range and accumulated on invading genotypes. Thus variation in microbial associations of invaders was driven by the interaction of G and E, and microbial communities appear to change in composition along with host evolution during invasion.

evolutionary biology↗

bioRxiv2023

Single-cell and spatial transcriptomics identify a macrophage population associated with skeletal muscle fibrosis

The monocytic/macrophage system is essential for skeletal muscle homeostasis, but its dysregulation contributes to the pathogenesis of muscle degenerative disorders. Despite our increasing knowledge of the role of macrophages in degenerative disease, it still remains unclear how macrophages contribute to muscle fibrosis. Here, we used single-cell transcriptomics to determine the molecular attributes of dystrophic and healthy muscle macrophages. We identified six novel clusters. Unexpectedly, none corresponded to traditional definitions of M1 or M2 macrophage activation. Rather, the predominant macrophage signature in dystrophic muscle was characterized by high expression of fibrotic factors, galectin-3 and spp1. Spatial transcriptomics and computational inferences of intercellular communication indicated that spp1 regulates stromal progenitor and macrophage interactions during muscular dystrophy. Galectin-3+ macrophages were chronically activated in dystrophic muscle and adoptive transfer assays showed that the galectin-3+ phenotype was the dominant molecular program induced within the dystrophic milieu. Histological examination of human muscle biopsies revealed that galectin-3+ macrophages were also elevated in multiple myopathies. These studies advance our understanding of macrophages in muscular dystrophy by defining the transcriptional programs induced in muscle macrophages, and reveal spp1 as a major regulator of macrophage and stromal progenitor interactions.

immunology↗

bioRxiv2023

PhageDPO: Phage Depolymerase Finder

Bacteriophages are the most predominant and genetically diverse biological entities on Earth. They are bacterial viruses which encode numerous proteins with potential antibacterial activity. However, most bacteriophage-encoded proteins have no assigned function, hindering the discovery of novel antibacterial agents. In particular, there has been a growing interest in exploring recombinant bacteriophage depolymerases from the fundamental standpoint, but mostly for biotechnological applications to control bacterial pathogens. Due to the lack of efficient identification tools, we developed PhageDPO, the first developed tool that predicts depolymerases in bacteriophage genomes using machine learning methods. Availability and implementationPhageDPO was integrated into a Galaxy framework available online at: bit.ly/phagedpo.

bioinformatics↗

bioRxiv2021

Prediction of standard cell types and functional markers from flow cytometry gating definitions using machine learning

BackgroundA key step in clinical flow cytometry data analysis is gating, which involves the identification of cell populations. The process of gating produces a set of reportable results, which are typically described by gating definitions. The non-standardized, non-interpreted nature of gating definitions represents a hurdle for data interpretation and data sharing across and within organizations. Interpreting and standardizing gating definitions for subsequent analysis of gating results requires a curation effort from experts. Machine learning approaches have the potential to help in this process by predicting expert annotations associated with gating definitions. MethodsWe created a gold-standard dataset by manually annotating thousands of gating definitions with cell type and functional marker annotations. We used this dataset to train and test a machine learning pipeline able to predict standard cell types and functional marker genes associated with gating definitions. ResultsThe machine learning pipeline predicted annotations with high accuracy for both cell types and functional marker genes. Accuracy was lower for gating definitions from assays belonging to laboratories from which limited or no prior data was available in the training. Manual error review ensured that resulting predicted annotations could be reused subsequently as additional gold-standard training data. ConclusionsMachine learning methods are able to consistently predict annotations associated with gating definitions from flow cytometry assays. However, a hybrid automatic and manual annotation workflow would be recommended to achieve optimal results.

molecular biology↗

bioRxiv2018

High throughput genotyping of structural variations in a complex plant genome using an original Affymetrix(R) Axiom(R) array

BackgroundInsertions/deletions (InDels) and more specifically presence/absence variations (PAVs) are pervasive in several species and have strong functional and phenotypic effect by removing or drastically modifying genes. Genotyping of such variants on large panels remains poorly addressed, while necessary for approaches such as association mapping or genomic selection.\n\nResultsWe have developed, as a proof of concept, a new high-throughput and affordable approach to genotype InDels. We first identified 141,000 InDels by aligning reads from the B73 line against the genome of three temperate maize inbred lines (F2, PH207, and C103) and reciprocally. Next, we designed an Affymetrix(R) Axiom(R) array to target these InDels, with a combination of probes selected at breakpoint sites (13%) or within the InDel sequence, either at polymorphic (25%) or non-polymorphic sites (63%) sites. The final array design is composed of 662,772 probes and targets 105,927 InDels, including PAVs ranging from 35bp to 129kbp. After Affymetrix(R) quality control, we successfully genotyped 86,648 polymorphic InDels (82% of all InDels interrogated by the array) on 445 maize DNA samples with 422,369 probes. Genotyping InDels using this approach produced a highly reliable dataset, with low genotyping error (~3%), high call rate (~98%), and high reproducibility (>95%). This reliability can be further increased by combining genotyping of several probes calling the same InDels (<0.1% error rate and >99.9% of call rate for 5 probes). This \"proof of concept\" tool was used to estimate the kinship matrix between 362 maize lines with 57,824 polymorphic InDels. This InDels kinship matrix was highly correlated with kinship estimated using SNPs from Illumina 50K SNP arrays.\n\nConclusionsWe efficiently genotyped thousands of small to large InDels on a sizeable number of individuals using a new Affymetrix(R) Axiom(R) array. This powerful approach opens the way to studying the contribution of InDels to trait variation and heterosis in maize. The approach is easily extendable to other species and should contribute to decipher the biological impact of InDels at a larger scale.

genomics↗

bioRxiv2017

Optimal homotopy analysis of a chaotic HIV-1 model incorporating AIDS-related cancer cells

The studies of nonlinear models in epidemiology have generated a deep interest in gaining insight into the mechanisms that underlie AIDS-related cancers, providing us with a better understanding of cancer immunity and viral oncogenesis. In this article, we analyse an HIV-1 model incorporating the relations between three dynamical variables: cancer cells, healthy CD4+ T lymphocytes and infected CD4+ T lymphocytes. Recent theoretical investigations indicate that these cells interactions lead to different dynamical outcomes, for instance to periodic or chaotic behavior. Firstly, we analytically prove the boundedness of the trajectories in the systems attractor. The complexity of the coupling between the dynamical variables is quantified using observability indices. Our calculations reveal that the highest observable variable is the population of cancer cells, thus indicating that these cells could be monitored in future experiments in order to obtain time series for attractors reconstruction. We identify different dynamical behaviors of the system varying two biologically meaningful parameters: r1, representing the uncontrolled proliferation rate of cancer cells, and k1, denoting the immune systems killing rate of cancer cells. The maximum Lyapunov exponent is computed to identify the chaotic regimes. Considering very recent developments in the literature related to the homotopy analysis method (HAM), we construct the explicit series solution of the cancer model and focus our analysis on the dynamical variable with the highest observability index. An optimal homotopy analysis approach is used to improve the computational efficiency of HAM by means of appropriate values for the convergence control parameter, which greatly accelerate the convergence of the series solution.

systems biology↗