Biology subjects

Becerra, D.

Publications and source records attributed to Becerra, D..

4 recordsLinked to original sources

bioRxiv2026

Oral administration of dibenzoylmethane (DBM) prevents cognitive decline in a C9ORF72-mediated FTD mouse model

Amyotrophic lateral sclerosis (ALS) and frontotemporal dementia (FTD) are two related neurodegenerative disorders that display overlapping features. The hexanucleotide repeat expansion GGGGCC (G4C2) in the C9ORF72 gene is the most common cause of ALS and FTD, which results in the accumulation of dipeptide-repeat protein aggregates. Regulation of protein synthesis at the level of the initiation factor eIF2 has been suggested as a transversal event contributing to neurodegeneration in ALS and FTD. eIF2 phosphorylation blocks protein synthesis to alleviate protein misfolding overload, but conversely it can reduce the expression of synaptic proteins resulting in neuronal dysfunction. Dibenzoylmethane (DBM) is a small molecule that reverses the translational attenuation mediated by eIF2 phosphorylation which has been shown to alleviate neurodegeneration in prion-infected mice and Tau transgenic animals. Here we investigated the efficacy of the oral administration of DBM in protecting a mouse model of C9ORF72 pathogenesis. Treatment of mice with 0.5% of DBM mixture in powdered food ad libitum was sufficient to prevent cognitive impairment in C9ORF72 mice. Unexpectedly, DBM treatment did not modify the content of poly(GA) and poly(GR) protein inclusion in the hippocampus and brain cortex. Proteomic profiling of brain tissue indicated that DBM administration corrected nearly 70% of the changes in gene expression triggered by expanded G4C2, where the main pathways modified by DBM were related to cytoskeleton organization, ALS, and metabolic processes. Most proteins corrected by DBM in our C9ORF72 model were also altered in the brain of human FTD/ALS patients. Overall, our results reinforce the idea that targeting protein synthesis with small molecules in patients carrying C9ORF72 mutations may result in improved cognitive capacity.

molecular biology↗

bioRxiv2025

Noradrenaline and Acetylcholine shape Functional Connectivity organization of NREM substages: an empirical and simulation study

Sleep onset is characterized by a departure from arousal, and can be separated into well-differentiated stages: NREM (which encompasses three substages: N1, N2 and N3) and REM (Rapid Eye Movement). Awake brain dynamics are maintained by various wake-promoting mechanisms, particularly the neuromodulators Acetylcholine (ACh) and Noradrenaline (NA), whose levels naturally decrease during the transition to sleep. The combined influence of these neurotransmitters on brain connectivity during sleep remains unclear, as previous models have examined them mostly in isolation or only in deep sleep. In this study, we employ a Whole Brain model to investigate how changes in brain neurochemistry during NREM sleep, specifically involving ACh and NA, affect the Functional Connectivity (FC) of the brain. Using a Wilson-Cowan whole brain model informed by an empirical connectome and a heterogeneous receptivity map of neuromodulators, we explore these dynamics. Initial FC analysis reveals distinct connectivity changes: a decrease in Locus Coeruleus (LC) connectivity with the cortex during N2 and N3, and a decrease in Basal Forebrain (BF) connectivity with the cortex during N3. Additionally, compared to Wakefulness (W), there is a transition to a more integrated state in N1 and a more segregated state in N3. We adjust the coupling and input-output slope of the Whole Brain model for ACh and NA, based on BF or LC priors, to show that region-specific neurotransmitter distribution is key to explaining their effects on FC. This work enhances our understanding of neurotransmitters roles in modulating sleep stages and their significant contribution to brain state transitions between different states of consciousness, both in health and disease.

neuroscience↗

bioRxiv2023

Semi-supervised learning and large-scale docking data accelerate RNA virtual screening

RNAs constitute a vast reservoir of mostly untapped drug targets. Structure-based virtual screening (VS) methods screen large compound libraries for identifying promising candidate molecules by conditioning on binding site information. The classical approach relies on molecular docking simulations. However, this strategy does not scale well with the size of the small molecule databases and the number of potential RNA targets. Machine learning emerged as a promising technology to resolve this bottleneck. Efficient data-driven VS methods have already been introduced for proteins, but these techniques have not yet been developed for RNAs due to limited dataset sizes and lack of practical use-case evaluation. We propose a data-driven VS pipeline that deals with the unique challenges of RNA molecules through coarse grained modeling of 3D structures and heterogeneous training regimes using synthetic data augmentation and RNA-centric self supervision. We report strong prediction and generalizability of our framework, ranking active compounds among inactives in the top 2.8% on average on a structurally distinct drug-like test set. Those predictions are sensitive, but robust to pockets alterations, opening the door to its use on binding site detection methods outputs. Our model results in a ten thousand-times speedup over docking techniques while obtaining higher performance. Finally, we deploy our model on a recently published in-vitro small molecule microarray experiment with 20,000 compounds and report a mean enrichment factor at 1% of 2.93 on four unseen RNA riboswitches. To our knowledge, this is the first experimental evidence of success for structure-based deep learning methods in RNA virtual screening. Our source code and data, as well as a Google Colab notebook for inference, are available on GitHub.1

bioinformatics↗

bioRxiv2019

Designing real novel proteins using deep graph neural networks

Protein structure and function is determined by the arrangement of the linear sequence of amino acids in 3D space. Despite substantial advances, precisely designing sequences that fold into a predetermined shape (the "protein design" problem) remains difficult. We show that a deep graph neural network, ProteinSolver, can solve protein design by phrasing it as a constraint satisfaction problem (CSP). To sidestep the considerable issue of optimizing the network architecture, we first develop a network that is accurately able to solve the related and straightforward problem of Sudoku puzzles. Recognizing that each protein design CSP has many solutions, we train this network on millions of real protein sequences corresponding to thousands of protein structures. We show that our method rapidly designs novel protein sequences and perform a variety of in silico and in vitro validations suggesting that our designed proteins adopt the predetermined structures. One Sentence SummaryA neural network optimized using Sudoku puzzles designs protein sequences that adopt predetermined structures.

bioinformatics↗