Biology subjects

Fonseca, G. J.

Publications and source records attributed to Fonseca, G. J..

7 recordsLinked to original sources

bioRxiv2025

AmalgaMo: flexible DNA motif merging

MotivationInference of candidate upstream regulators via motif enrichment analysis is a common step in the interpretation of genomic data. However, redundancy in motif databases can negatively impact predictive value, especially when relying on regression-based motif enrichment analysis. Although various forms of motif clustering have been used to mitigate problems caused by redundancy, an algorithm optimized for downstream regression-based analysis is needed. ResultsWe introduce AmalgaMo, an efficient and flexible command line tool for merging highly similar motifs. Using publicly available human datasets, we demonstrate that merging motifs with our optimized settings greatly benefits regression-based motif enrichment analysis and provide detailed documentation that can serve as a reference for researchers inferring upstream regulators from genomic data. AvailabilitySource code is available on GitHub at https://github.com/lapohosorsolya/AmalgaMo.

genomics↗

bioRxiv2025

Refining sequence-to-expression modelling with chromatin accessibility

MotivationSequence-to-expression models typically do not consider chromatin accessibility, a major factor limiting gene regulation. We hypothesized that supplying accessibility as an input feature would allow a sequenceto-expression model to focus on important open regions of the genome. ResultsWe found that the performance of such an augmented model was significantly better than that of sequence-only or accessibility-only models with similar architectures. Specifically, its ability to predict the expression of highly variable genes and gene expression in other cell types improved, and higher attribution scores in the input DNA sequences of the augmented model conformed to accessibility, enabling the learning of cell type-specific sequence patterns. Additionally, we show that fine-tuning a pre-trained sequence-only model with both sequence and accessibility can boost performance further and highlight the importance of sequencing depth in sequence-toexpression prediction. Availability and ImplementationSource code is available on GitHub at https://github.com/lapohosorsolya/accessible_seq2exp.

bioinformatics↗

bioRxiv2023

scSemiProfiler: Advancing Large-scale Single-cell Studiesthrough Semi-profiling with Deep Generative Models andActive Learning

AbstractSingle-cell sequencing is a crucial tool for dissecting the cellular intricacies of complex diseases. Its prohibitive cost, however, hampers its application in expansive biomedical studies. Traditional cellular deconvolution approaches can infer cell type proportions from more affordable bulk sequencing data, yet they fall short in providing the detailed resolution required for single-cell-level analyses. To overcome this challenge, we introduce "scSemiProfiler", an innovative computational framework that marries deep generative models with active learning strategies. This method adeptly infers single-cell profiles across large cohorts by fusing bulk sequencing data with targeted single-cell sequencing from a few rigorously chosen representatives. Extensive validation across heterogeneous datasets verifies the precision of our semi-profiling approach, aligning closely with true single-cell profiling data and empowering refined cellular analyses. Originally developed for extensive disease cohorts, "scSemiProfiler" is adaptable for broad applications. It provides a scalable, cost-effective solution for single-cell profiling, facilitating in-depth cellular investigation in various biological domains.

bioinformatics↗

bioRxiv2022

Machine-guided cell-fate engineering

AO_SCPLOWBSTRACTC_SCPLOWThe creation of induced pluripotent stem cells (iPSCs) has enabled scientists to explore the derivation of many types of cells. While there are diverse general approaches for cell-fate engineering, one of the fastest and most efficient approaches is transcription factor (TF) over-expression. However, finding the right combination of TFs to over-express to differentiate iPSCs directly into other cell-types is a difficult task. Here were describe a machine-learning (ML) pipeline, called CellCartographer, for using chromatin accessibility data to design multiplex TF pooled-screens for cell type conversions. We validate this method by differentiating iPSCs into twelve diverse cell types at low efficiency in preliminary screens and then iteratively refining our TF combinations to achieve high efficiency differentiation for six of these cell types in < 6 days. Finally, we functionally characterized engineered iPSC-derived cytotoxic T-cells (iCytoT), regulatory T-cells (iTReg), type II astrocytes (iAstII), and hepatocytes (iHep) to validate functionally accurate differentiation.

developmental biology↗

bioRxiv2021

Direct reprogramming of the intestinal epithelium by parasitic helminths subverts type 2 immunity

Enteric helminths form intimate physical connections with the intestinal epithelium, yet their ability to directly alter epithelial stem cell fate has not been resolved. Here we demonstrate that infection of mice with the symbiotic parasite Heligmosomoides polygyrus bakeri (Hbp), reprograms the intestinal epithelium into a fetal-like state marked by the emergence of Clusterin-expressing revival stem cells (revSCs). Organoid-based studies using parasite-derived excretory/secretory products reveal that Hpb-mediated revSC generation occurs independent of host-derived immune signals and inhibits type 2 cytokine-driven differentiation of secretory epithelial lineages that promote their expulsion. Reciprocally, type 2 cytokine signals limit revSC differentiation and, consequently, Hpb fitness indicating that helminths compete with their host for control of the intestinal stem cell niche to promote continuation of their life cycle.

immunology↗

bioRxiv2020

Natural genetic variation affecting transcription factor spacing at regulatory regions is generally well tolerated

Regulation of gene expression requires the combinatorial binding of sequence-specific transcription factors (TFs) at promoters and enhancers. Prior studies showed that alterations in the spacing between TF binding sites can influence promoter and enhancer activity. However, the relative importance of TF spacing alterations resulting from naturally occurring insertions and deletions (InDels) has not been systematically analyzed. To address this question, we first characterized the genome-wide spacing relationships of 75 TFs in K562 cells as determined by ChIP-sequencing. We found a dominant pattern of a relaxed range of spacing between collaborative factors, including 46 TFs exclusively exhibiting relaxed spacing with their binding partners. Next, we exploited millions of InDels provided by genetically diverse mouse strains and human individuals to investigate the effects of altered spacing on TF binding and local histone acetylation. Spacing alterations resulting from naturally occurring InDels are generally tolerated in comparison to genetic variants directly affecting TF binding sites. A remarkable range of tolerance was further established for PU.1 and C/EBP{beta}, which exhibit relaxed spacing, by introducing synthetic spacing alterations ranging from 5-bp increase to >30-bp decrease using CRISPR/Cas9 mutagenesis. These findings provide implications for understanding mechanisms underlying enhancer selection and for the interpretation of non-coding genetic variation.

genomics↗

bioRxiv2018

Diverse motif ensembles specify non-redundant DNA binding activities of AP-1 family members in macrophages

Mechanisms by which members of the AP-1 family of transcription factors play both redundant and non-redundant biological roles despite recognizing the same DNA sequence remain poorly understood. To address this question, we investigated the molecular functions and genome-wide DNA binding patterns of AP-1 family members in macrophages. ChIP-sequencing showed overlapping and distinct binding profiles for each factor that were remodeled following TLR4 ligation. Development of a machine learning approach that jointly weighs hundreds of DNA recognition elements yielded dozens of motifs predicted to drive factor-specific binding profiles. Machine learning-based predictions were confirmed by analysis of the effects of mutations in genetically diverse mice and by loss of function experiments. These findings provide evidence that non-redundant genomic locations of different AP-1 family members in macrophages largely result from collaborative interactions with diverse, locus-specific ensembles of transcription factors and suggest a general mechanism for encoding functional specificities of their common recognition motif.

genomics↗