Biology subjects

Fornes, O.

Publications and source records attributed to Fornes, O..

16 recordsLinked to original sources

bioRxiv2026

Deep learning design and in vivo validation of Müller glia-specific cis-regulatory elements

Effective recombinant adeno-associated virus gene therapies require promoters that are compact and cell-type-specific. Ideally, promoters should also exhibit functional conservation when tested in model organisms to ensure that preclinical findings translate reliably to human patients. Here, we introduce a deep learning framework for designing cis-regulatory elements (CREs) meeting these criteria, applied to retinal Muller glia (MG). Using single-cell chromatin accessibility data from human and mouse retinas, we trained species-specific models to predict cell-type accessibility, and designed compact CREs using two complementary strategies. In silico validation predicted that the designed CREs exhibit high MG-specific accessibility (on-target) in both species with minimal off-target accessibility across hundreds of human cell types and tissues. Mechanistic analysis revealed that the predicted MG accessibility is driven by the creation of LHX2 motifs. In vivo validation confirmed that the designed CREs successfully restrict reporter expression to MG in the murine retina. Our deep learning framework is highly generalizable and enables the rapid design of compact, on-target, and species-conserved CREs for precision gene therapy.

genomics↗

bioRxiv2026

ModCRElib: A standalone package to model cis-regulatory elements.

SummaryThe ModCRElib package provides various tools for the analysis and modelling of transcription factor(TF)-DNA and regulatory complex inter-protein interactions. It takes structural information on these interactions to predict TF binding motifs, generate binding profiles along DNA sequences that score the binding affinity, predict TF binding sites and model the structure of higher order regulatory complexes. It is capable of working with a variety of input data formats and sources. The user may follow the analysis pipeline as outlined in the documentation or the user can make use of any of the multiple functionalities in an isolated manner. The package takes the service offered by the ModCRE server and enables users to apply its tools in an unrestricted and customizable manner. In this paper we provide 5 example uses of ModCRElib. This includes (i) TF binding affinity prediction, (ii) TF binding aggregation, (iii) characterization of specificity in TF binding sites along target DNA sequences,(iv) modelling TF bound to predicted binding sites, and (v) the generation of statistical potential derived scoring profiles of TF interacting with DNA. Availabilityhttps://github.com/structuralbioinformatics/ModCRElib Contactbaldo.oliva@upf.edu Supplementary informationAvailable at https://github.com/structuralbioinformatics/ModCRElib. doi:10.5281/zenodo.17484081

bioinformatics↗

bioRxiv2025

Inferring binding specificities of human transcription factors with the wisdom of crowds

DNA motif discovery and, particularly, computational modeling of transcription factor binding motifs, has been a mecca of algorithmic bioinformatics for several decades. Here, we report the results of the largest open community challenge in Inferring BInding Specificities (IBIS), where participants all over the world were invited to construct binding specificity models from multi-assay experimental data for poorly studied human transcription factors. The submissions were rigorously tested against a rich held-out dataset. Benchmarking demonstrated a consistent advantage of properly designed deep learning models over traditional positional weight matrices and other machine learning methods. Yet, the positional weight matrices displayed a surprisingly strong performance out of the box, being only slightly behind the best deep learning models. A post-challenge assessment of a selection of other deep learning methods further solidified this finding. IBIS highlights the power of benchmarking in finding adequate DNA motif representations, emphasizes the pros and cons of various machine learning methods applied to DNA motif modeling, and establishes a rich dataset, benchmarking protocols, and computational framework for a fair cross-platform evaluation of future models of transcription factor binding motifs in DNA sequences. Graphical Abstract O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=175 SRC="FIGDIR/small/688692v1_ufig1.gif" ALT="Figure 1"> View larger version (64K): org.highwire.dtl.DTLVardef@1c6677corg.highwire.dtl.DTLVardef@b4124aorg.highwire.dtl.DTLVardef@1ce2b1org.highwire.dtl.DTLVardef@66e917_HPS_FORMAT_FIGEXP M_FIG C_FIG

bioinformatics↗

bioRxiv2024

Cross-platform DNA motif discovery and benchmarking to explore binding specificities of poorly studied human transcription factors

A DNA sequence pattern, or "motif", is an essential representation of DNA-binding specificity of a transcription factor (TF). Any particular motif model has potential flaws due to shortcomings of the underlying experimental data and computational motif discovery algorithm. As a part of the Codebook/GRECO-BIT initiative, here we evaluated at large scale the cross-platform recognition performance of positional weight matrices (PWMs), which remain popular motif models in many practical applications. We applied ten different DNA motif discovery tools to generate PWMs from the "Codebook" data comprised of 4,237 experiments from five different platforms profiling the DNA-binding specificity of 394 human proteins, focusing on understudied transcription factors of different structural families. For many of the proteins, there was no prior knowledge of a genuine motif. By benchmarking-supported human curation, we constructed an approved subset of experiments comprising about 30% of all experiments and 50% of tested TFs which displayed consistent motifs across platforms and replicates. We present the Codebook Motif Explorer (https://mex.autosome.org), a detailed online catalog of DNA motifs, including the top-ranked PWMs, and the underlying source and benchmarking data. We demonstrate that in the case of high-quality experimental data, most of the popular motif discovery tools detect valid motifs and generate PWMs, which perform well both on genomic and synthetic data. Yet, for each of the algorithms, there were problematic combinations of proteins and platforms, and the basic motif properties such as nucleotide composition and information content offered little help in detecting such pitfalls. By combining multiple PMWs in decision trees, we demonstrate how our setup can be readily adapted to train and test binding specificity models more complex than PWMs. Overall, our study provides a rich motif catalog as a solid baseline for advanced models and highlights the power of the multi-platform multi-tool approach for reliable mapping of DNA binding specificities. O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=141 SRC="FIGDIR/small/619379v2_ufig1.gif" ALT="Figure 1"> View larger version (61K): org.highwire.dtl.DTLVardef@79561forg.highwire.dtl.DTLVardef@54c0aorg.highwire.dtl.DTLVardef@1c33f34org.highwire.dtl.DTLVardef@16a93ba_HPS_FORMAT_FIGEXP M_FIG O_FLOATNOGraphical AbstractC_FLOATNO C_FIG

bioinformatics↗

bioRxiv2024

Perspectives on Codebook: sequence specificity of uncharacterized human transcription factors

Gene expression is regulated by transcription factors (TFs), which recognize specific DNA sequence motifs. Several hundred putative human TFs, identified mainly by an apparent DNA-binding domain, lack known binding motifs1, and even for well-characterized TFs, it remains controversial to what degree motifs accurately reflect binding sites in living cells2,3. Here, we describe a systematic effort ("Codebook") to determine the sequence specificity of 332 putative and poorly characterized human TFs. Over 4,000 independent experiments, encompassing multiple in vitro and in vivo assays, produced motifs for just over half (177, or 53%), of which most are unique to a single protein, thereby extending the vocabulary of sequence recognition encoded by human TFs by [~]100 distinct motifs. Moreover, binding motifs identified in vitro are strongly enriched within cellular binding sites. Collectively, the data reveal tens of thousands of previously unknown, conserved, and direct TF binding sites across the human genome. These sites are concentrated in promoter regions, and are predictive of gene expression, illustrating that this new data atlas provides an important step forward in decoding the human genome.

genomics↗

bioRxiv2022

A new type of transcriptional reprogramming by an IRF4 mutation in lymphoma

SUMMARY PARAGRAPHDisease-causing mutations in genes encoding transcription factors (TFs) are a recurrent finding in hematopoietic malignancies and might involve key regulators of lineage adherence and cellular differentiation1-3. Such mutations can affect TF-interactions with their cognate DNA-binding motifs4, 5. Whether and how TF-mutations impact upon the nature of binding to TF composite elements (CE) and influence their interaction with other TFs is unclear. Here, we report a new mechanism of TF alteration in human lymphomas with perturbed B cell identity. It is caused by a recurrent somatic missense mutation c.295T>C (p.Cys99Arg; p.C99R) targeting the center of the DNA-binding domain of Interferon Regulatory Factor 4 (IRF4), a key TF in immune cell-differentiation and -activation6, 7. IRF4-C99R fundamentally alters IRF4 DNA-binding, with loss-of-binding to canonical IRF motifs and neomorphic gain-of-binding to canonical and non-canonical IRF composite elements (CEs). Furthermore, IRF4-C99R thoroughly modifies IRF4 function, by blocking IRF4-dependent plasma cell induction, and up-regulating disease-specific genes in a non-canonical Activator Protein-1 (AP-1)-IRF-CE (AICE)-dependent manner. Our data explain how a single arginine mutation creates a complex switch of TF specificity and gene regulation. These data open the possibility of designing specific inhibitors to block the neomorphic, disease-causing DNA-binding activities of a mutant transcription factor.

cancer biology↗

bioRxiv2022

ExplaiNN: interpretable and transparent neural networks for genomics

Sequence-based deep learning models, particularly convolutional neural networks (CNNs), have shown superior performance on a wide range of genomic tasks. A key limitation of these models is the lack of interpretability, slowing down their adoption by the genomics community. Current approaches to model interpretation do not readily reveal how a model makes predictions, can be computationally intensive, and depend on the implemented architecture. Here, we introduce ExplaiNN, an adaptation of neural additive models[1] for genomic tasks wherein predictions are computed as a linear combination of multiple independent CNNs, each consisting of a single convolutional filter and fully connected layers. This approach brings together the expressiveness of CNNs with the interpretability of linear models, providing global (cell state level) as well as local (individual sequence level) biological insights into the data. We use ExplaiNN to predict transcription factor (TF) binding and chromatin accessibility states, demonstrating performance levels comparable to state-of-the-art methods, while providing a transparent view of the models predictions in a straightforward manner. Applied to de novo motif discovery, ExplaiNN identifies equivalent motifs to those obtained from specialized algorithms across a range of datasets. Finally, we present ExplaiNN as a plug-and-play platform in which pretrained TF binding models and annotated position weight matrices from reference databases can be easily combined. We expect that ExplaiNN will accelerate the adoption of deep learning by biological domain experts in their daily genomic sequence analyses.

genomics↗

bioRxiv2022

ModCRE: a structure homology-modeling approach to predict TF binding in cis-regulatory elements

Transcription factor (TF) binding is a key component of genomic regulation. There are numerous high-throughput experimental methods to characterize TF-DNA binding specificities. Their application, however, is both laborious and expensive, which makes profiling all TFs challenging. For instance, the binding preferences of ~25% human TFs remain unknown; they neither have been determined experimentally nor inferred computationally. We introduce a structure-based learning approach to predict the binding preferences of TFs and the automated modelling of TF regulatory complexes. We show the advantage of using our approach over the state-of-art nearest-neighbor prediction in the limits of remote homology. Starting from a TF sequence or structure, we predict binding preferences in the form of motifs that are then used to scan a DNA sequence for occurrences. The best matches are either profiled with a binding score or collected for their subsequent modeling into a higher-order regulatory complex with DNA. Cooperativity is modelled by: i) the co-localization of TFs; and ii) the structural modeling of protein-protein interactions between TFs and with co-factors. As case examples, we apply our approach to automatically model the interferon-{beta} enhanceosome and the pioneering complex of OCT4, SOX2 and SOX11 with a nucleosome, which are compared with the experimentally known structures.

bioinformatics↗

bioRxiv2021

In silico discovery of small molecules for efficient stem cell differentiation into definitive endoderm

Improving methods for human embryonic stem cell differentiation represents a challenge in modern regenerative medicine research. Using drug repurposing approaches, we discover small molecules that regulate the formation of definitive endoderm. Among them are inhibitors of known processes involved in endoderm differentiation (mTOR, PI3K, and JNK pathways) and a new compound, with an unknown mechanism of action, capable of inducing endoderm formation in the absence of growth factors in the media. Optimization of the classical protocol by including this compound achieves the same differentiation efficiency with a 90% cost reduction. The gene expression profile induced by the compound suggests that it is an inhibitor of the MYC pathway. The proposed in silico procedure for candidate molecule selection has broad potential for improving stem cell differentiation protocols.

bioinformatics↗

bioRxiv2020

Biologically-relevant transfer learning improves transcription factor binding prediction

BackgroundDeep learning has proven to be a powerful technique for transcription factor (TF) binding prediction, but requires large training datasets. Transfer learning can reduce the amount of data required for deep learning, while improving overall model performance, compared to training a separate model for each new task. ResultsWe assess a transfer learning strategy for TF binding prediction consisting of a pre-training step, wherein we train a multi-task model with multiple TFs, and a fine-tuning step, wherein we initialize single-task models for individual TFs with the weights learned by the multi-task model, after which the single-task models are trained at a lower learning rate. We corroborate that transfer learning improves model performance, especially if in the pre-training step the multi-task model is trained with biologically-relevant TFs. We show the effectiveness of transfer learning for TFs with [~]500 ChIP-seq peak regions. Using model interpretation techniques, we demonstrate that the features learned in the pre-training step are refined in the fine-tuning step to resemble the binding motif of the target TF (i.e. the recipient of transfer learning in the fine-tuning step). Moreover, pre-training with biologically-relevant TFs allows single-task models in the fine-tuning step to learn features other than the motif of the target TF. ConclusionsOur results confirm that transfer learning is a powerful technique for TF binding prediction.

bioinformatics↗

bioRxiv2020

Cross-species examination of X-chromosome inactivation highlights domains of escape from silencing

BackgroundX-chromosome inactivation (XCI) in eutherian mammals is the epigenetic inactivation of one of the two X chromosomes in XX females in order to compensate for dosage differences with XY males. Not all genes are inactivated, and the proportion escaping from inactivation varies between human and mouse (the two species that have been extensively studied). ResultsWe used DNA methylation to predict the XCI status of X-linked genes with CpG islands across 12 different species: human, chimp, bonobo, gorilla, orangutan, mouse, cow, sheep, goat, pig, horse and dog. We determined the XCI status of 342 CpG islands on average per species, with most species having 80-90% of genes subject to XCI. Mouse was an outlier, with a higher proportion of genes subject to XCI than found in other species. Sixteen genes were found to have discordant X-chromosome inactivation statuses across multiple species, with five of these showing primate-specific escape from XCI. These discordant genes tended to cluster together within the X chromosome, along with genes with similar patterns of escape from XCI. CTCF- binding, ATAC-seq signal and LTR repeats were enriched at genes escaping XCI when compared to genes subject to XCI; however, enrichment was only observed in three or four of the species tested. LINE and DNA repeats showed enrichment around subject genes, but again not in a consistent subset of species. ConclusionsIn this study we determined XCI status across 12 species, showing mouse to be an outlier with few genes that escape inactivation. Inactivation status is largely conserved across species. The clustering of genes that change XCI status across species implicates a domain-level control. In contrast, the relatively consistent, but not universal correlation of inactivation status with enrichment of repetitive elements or CTCF binding at promoters demonstrates gene-based influences on inactivation state. This study broadens enrichment analysis of regulatory elements to species beyond human and mouse.

genetics↗

bioRxiv2020

A GO catalogue of human DNA-binding transcription factors

DNA-binding transcription factors recognise genomic addresses, specific sequence motifs in gene regulatory regions, to control gene transcription. A complete and reliable catalogue of all DNA-binding transcription factors is key to investigating the delicate balance of gene regulation in response to environmental and developmental stimuli. The need for such a catalogue of proteins is demonstrated by the many lists of DNA-binding transcription factors that have been produced over the past decade. The COST Action Gene Regulation Ensemble Effort for the Knowledge Commons (GREEKC) Consortium brought together experts in the field of transcription with the aim of providing high quality and interoperable gene regulatory data. The Gene Ontology (GO) Consortium provides strict definitions for gene product function, including factors that regulate transcription. The collaboration between the GREEKC and GO Consortia has enabled the application of those definitions to produce a new curated catalogue of human DNA-binding transcription factors, that can be accessed at https://www.ebi.ac.uk/QuickGO/targetset/dbTF. In addition, this curation effort has led to the GO annotation of almost sixty thousand DNA-binding transcription factors in over a hundred species. Thus, this work will aid researchers investigating the regulation of transcription in both biomedical and basic science.

bioinformatics↗

bioRxiv2020

GeneBreaker: Variant simulation to improve the diagnosis of Mendelian rare genetic diseases

Mendelian rare genetic diseases affect 5-10% of the population, and with over 5,300 genes responsible for ~7,000 different diseases, they are challenging to diagnose. The use of whole genome sequencing (WGS) has bolstered the diagnosis rate significantly. Effective use of WGS relies upon the ability to identify the disrupted gene responsible for disease phenotypes. This process involves genomic variant calling and prioritization, and is the beneficiary of improvements to sequencing technology, variant calling approaches, and increased capacity to prioritize genomic variants with potential pathogenicity. As analysis pipelines continue to improve, careful testing of their efficacy is paramount. However, real-life cases typically emerge anecdotally, and utilization of clinically sensitive and identifiable data for testing pipeline improvements is regulated and limiting. We identified the need for a gene-based variant simulation framework which can create mock rare disease scenarios, utilizing known pathogenic variants or through the creation of novel gene-disrupting variants. To fill this need, we present GeneBreaker, a tool which creates synthetic rare disease cases with utility for benchmarking variant calling approaches, testing the efficacy of variant prioritization, and as an educational mechanism for training diagnostic practitioners in the expanding field of genomic medicine. GeneBreaker is freely available at http://GeneBreaker.cmmt.ubc.ca.

genomics↗

bioRxiv2019

On the prediction of DNA-binding preferences of C2H2-ZF domains using structural models: application on human CTCF.

Cis2-His2 zinc finger (C2H2-ZF) proteins are the largest family of transcription factors in human and higher metazoans. However, the DNA-binding preferences of many members of this family remain unknown. We have developed a computational method to predict these DNA-binding preferences. We combine information from crystal structures composed by C2H2-ZF domains and from bacterial one-hybrid experiments to compute scores for protein-DNA binding based on statistical potentials. We apply the scores to compute theoretical position weight matrices (PWMs) of proteins with a DNA-binding domain composed by C2H2-ZF domains, with the only requirement of an input structure (experimentally determined or modelled). We have tested the capacity to predict PWMs of zinc finger domains, successfully predicting 3-2 nucleotides of a trinucleotide binding site for about 70% variants of single zinc-finger domains of Zif268. We have also tested the capacity to predict the PWMs of proteins composed by three C2H2-ZF domains, successfully matching between 60% and 90% of the binding-site motif according to the JASPAR database. The tests are used as a proof of the capacity to scan a DNA fragment and find the potential binding sites of transcription-factors formed by C2H2-ZF domains. As an example, we have tested the approach to predict the DNA-binding preferences of the human chromatin binding factor CTCF.

bioinformatics↗

bioRxiv2018

RADI (Reduced Alphabet Direct Information): Improving execution time for direct-coupling analysis

MotivationDirect-coupling analysis (DCA) for studying the coevolution of residues in proteins has been widely used to predict the three-dimensional structure of a protein from its sequence. Current algorithms for DCA, although efficient, have a high computational cost of determining Direct Information (DI) values for large proteins or domains. In this paper, we present RADI (Reduced Alphabet Direct Information), a variation of the original DCA algorithm that simplifies the computation of DI values by grouping physicochemically equivalent residues.\n\nResultsWe have compared the first top ranking 40 pairs of DI values and their closest paired contact in 3D. The ranking is also compared with results obtained using a similar but faster approach based on Mutual Information (MI). When we simplify the number of symbols used to describe a protein sequence to 9, RADI achieves similar results as the original DCA (i.e. with the classical alphabet of 21 symbols), while reducing the computation time around 30-fold on large proteins (with length around 1000 residues) and with higher accuracy than predictions based on MI. Interestingly, the simplification produced by grouping amino acids into only two groups (polar and non-polar) is still representative of the physicochemical nature that characterizes the protein structure, having a relevant and useful predictive value, while the computation time is reduced between 100 and 2500-fold.\n\nAvailabilityRADI is available at https://github.com/structuralbioinformatics/RADI\n\nContactbaldo.oliva@upf.edu\n\nSupplementary informationSupplementary data is available in the git repository.

bioinformatics↗

bioRxiv2017

Altered transcription factor binding events predict personalized gene expression and confer insight into functional cis-regulatory variants

Deciphering the functional roles of cis-regulatory variants is a critical challenge in genome analysis and interpretation. We hypothesize that altered transcription factor (TF) binding events are a central mechanism by which cis-regulatory variants impact gene expression. We present TF2Exp, the first gene-based framework (to our knowledge) to predict the impact of altered TF binding on personalized gene expression based on cis-regulatory variants. Using data from lymphoblastoid cell lines, TF2Exp models achieved suitable performance for 3,060 genes. Alterations within DNase I hypersensitive, CTCF-bound, and tissue-specific TF-bound regions were the greatest contributors to the models. Our cis-regulatory variant-based TF2Exp models performed as well as the state-of-the-art SNP-based models, both in cross-validation and external validation. In addition, unlike SNP-based models, our TF2Exp models have the unique advantages to evaluate impact of uncommon variants and distinguish the functional roles of variants in linkage disequilibrium, showing broader utility for future human genetic studies.

bioinformatics↗