Biology subjects

Abate, A.

Publications and source records attributed to Abate, A..

13 recordsLinked to original sources

bioRxiv2026

Zero-shot evaluations expose structured generalization limits in predictive models of the brain

Artificial neural networks (ANNs) can predict neural responses to natural images with remarkable accuracy, making them promising models of human vision. However, prediction scores alone reveal little about what computations these models have captured or how well they generalize beyond the images on which they are trained. Here, we introduce a principled zero-shot model evaluation framework in which ANN-based encoding models are frozen before being tested on not just held out images but entirely new datasets. Our strongest tests repurpose decades of cognitive neuroscience experiments as diagnostic benchmarks for identifying which computations current models capture and where they fail. To implement this framework we built encoding models of human category-selective regions (FFA, PPA, and EBA) using several pretrained ANNs and evaluated these fixed models across 7 independent fMRI datasets and 31 cognitive experiments from 10 published studies. This framework revealed three broad patterns. First, zero-shot evaluations exposed systematic limits to model predictivity that were largely hidden by standard within-dataset cross-validation. These limits were structured varying across brain regions and model classes and depending strongly on the similarity between the images used to build and test the models. Second, current models reproduced some, but not all, classical findings from cognitive neuroscience, providing a diagnosis of the computations that current models have yet to capture. Finally, models that performed well on prediction tests also tended to succeed at reproducing cognitive neuroscience findings, suggesting that prediction and explanation are closely linked. Unlike prediction scores however, cognitive tests help diagnose where and why models fail. Together, the zero-shot evaluation framework provides a scalable and unified approach for evaluating brain models across prediction and cognitive neuroscience tests, which can help us better understand what current models capture and what still remains to be explained.

neuroscience↗

bioRxiv2026

Updating ACC preclinical models: characterization of two new patient-derived cell lines

AdrenoCortical Carcinoma (ACC) is an aggressive, rare and heterogenous malignancy, that requires diverse preclinical models. For this reason, the development of new cell lines is pivotal. Here we describe the development and characterization of two of them, SMAC-2 and SMAC-3, established from surgical specimens of ACC patients. The characterization included their mutational profiling, the evaluation of steroidogenic enzymes expression, secretory activity and the expression of steroid hormone receptors. The proliferative ability of these cells within a zebrafish embryos xenograft was also evaluated. SMAC-2 originated from a metastatic EDP-M-treated ACC in a female patient with Cushing syndrome and hyperandrogenism, while SMAC-3 derived from a male patient with a mitotane-treated local recurrence, with no sign of hypercortisolism. TP53 was mutated in both lines. SMAC-2 cells were characterized by a pathogenic alteration on CTTNB1 gene and a deletion of CDKN2A gene, while SMAC-3 on MSH2 gene. Basal hormonal status analysis showed a cell model-specific fingerprint either in the hormonal secretion and gene and protein expression of steroid hormone receptors. SMAC-2 secreted high levels of cortisol. SMAC-3 secreted low basal level of cortisol. Mitotane displayed in both cell lines a low potency. Under the experimental conditions used, the xenografted area did not increase for both cell models. Experiments were carried out to study the stability of the two cell lines. SMAC-2 and SMAC-3 display unique molecular and functional features, expanding the repertoire of experimental ACC models and representing valuable tools for preclinical research alongside established cell lines.

pharmacology and toxicology↗

bioRxiv2025

Deterministic Cell Pairing with Simultaneous Microfluidic Merging and Sorting of Droplets

Cell-cell interactions drive immune activation, tissue repair, and stem cell fate, yet there are few methods that can create large numbers of pre-defined cell pairs to study cell crosstalk. Droplet microfluidics allows high-throughput compartmentalization of multiple cells, but random loading results in <1% of droplets containing the desired cell combinations. Here, we present a microfluidic device capable of deterministically isolating specific cell pairs using droplet merging and sorting ( merge-sorting). The system detects target combinations using fluorescence and triggers simultaneous electrocoalescence and dielectrophoretic sorting. Using fluorescent dye- loaded droplets, we achieved 98.6% purity of merged and sorted droplets. In experiments using cells stained with three distinct dyes, >90% of desired cell pairs were recovered--compared to fewer than 1% when using random Poisson loading. To demonstrate the utility of this platform for extended co-culture studies, we merged cells in an alginate solution with calcium chloride droplets, producing monodisperse alginate hydrogels in which 92% of the beads contained target cell pairs that maintained viability over 18 hours. Compared to selective merger, this approach physically isolates desired droplets, eliminating unmerged contaminants and enabling cleaner downstream workflows. The device allows off-chip pre-incubation of droplets before pairing, the merger of reagents for multi-step assays, and the isolation of desired droplet pairs -- capabilities not jointly accessible with existing approaches. In summary, merge-sort is a flexible platform to enrich specific combinations of droplets, cells, or particles for high-throughput studies of cell crosstalk.

bioengineering↗

bioRxiv2024

Dlk1 is a novel adrenocortical stem/progenitor cell marker that predicts malignancy in adrenocortical carcinoma

Disruption of processes involved in tissue development and homeostatic self-renewal is increasingly implicated in cancer initiation, progression, and recurrence. The adrenal cortex is a dynamic tissue that undergoes life-long turnover. Here, using genetic fate mapping and murine adrenocortical carcinoma (ACC) models, we have identified a population of adrenocortical stem cells that express delta-like non-canonical Notch ligand 1 (DLK1). These cells are active during development, near dormant postnatally but are re-expressed in ACC. In a study of over 200 human ACC samples, we have shown DLK1 expression is ubiquitous and is an independent prognostic marker of recurrence-free survival. Paradoxically, despite its progenitor role, spatial transcriptomic analysis has identified DLK1 expressing cell populations to have increased steroidogenic potential in human ACC, a finding also observed in four human and one murine ACC cell lines. Finally, the cleavable DLK1 ectodomain is measurable in patients serum and can discriminate between ACC and other adrenal pathologies with high sensitivity and specificity to aid in diagnosis and follow-up of ACC patients. These data demonstrate a prognostic role for DLK1 in ACC, detail its hierarchical expression in homeostasis and oncogenic transformation and propose a role for its use as a biomarker in this malignancy. Graphical abstract O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=160 SRC="FIGDIR/small/609117v1_ufig1.gif" ALT="Figure 1"> View larger version (51K): org.highwire.dtl.DTLVardef@12aba74org.highwire.dtl.DTLVardef@374daaorg.highwire.dtl.DTLVardef@cbdc37org.highwire.dtl.DTLVardef@1e6aeac_HPS_FORMAT_FIGEXP M_FIG C_FIG Statement of significanceThis study presents DLK1 as a novel biomarker in ACC with opportunities for use in the diagnosis, prognosis and longitudinal follow up of patients. DLK1, a marker of adrenocortical stem cells, is re-expressed in ACC, is measurable in patients serum and is associated with increased malignancy.

cancer biology↗

bioRxiv2023

Multiomic Single Cell Sequencing Identifies Stemlike Nature of Mixed Phenotype Acute Leukemia and Provides Novel Risk Stratification

Mixed phenotype acute leukemia (MPAL) is a leukemia whose biologic drivers are poorly understood, therapeutic strategy remains unclear, and prognosis is poor. We performed multiomic single cell (SC) profiling of 14 newly diagnosed adult MPAL patients to characterize the immunophenotypic, genetic, and transcriptional landscapes of MPAL. We show that neither genetic profile nor transcriptome reliably correlate with specific MPAL immunophenotypes. However, progressive acquisition of mutations is associated with increased expression of immunophenotypic markers of immaturity. Using SC transcriptional profiling, we find that MPAL blasts express a stem cell-like transcriptional profile distinct from other acute leukemias and indicative of high differentiation potential. Further, patients with the highest differentiation potential demonstrated inferior survival in our dataset. A gene set score, MPAL95, derived from genes highly enriched in this cohort, is applicable to bulk RNA sequencing data and was predictive of survival in an independent patient cohort, suggesting utility for clinical risk stratification.

cancer biology↗

bioRxiv2022

Microfluidics-free single-cell genomics with templated emulsification

Single-cell RNA sequencing is now a standard method used to reveal the molecular details of cellular heterogeneity, but current approaches have limitations on speed, scale, and ease of use that stem from the complex microfluidic devices or fluid handling steps required for sample processing. We, therefore, developed a method that does not require specialized microfluidic devices, expertise, or hardware. Our approach is based on particle-templated emulsification, which allows single-cell encapsulation and barcoding of cDNA in uniform droplet emulsions with only a vortexer. PIP-seq accommodates a wide range of emulsification formats, including microwell plates and large-volume conical tubes, enabling thousands of samples or millions of cells to be processed in minutes. We demonstrate that PIP-seq produces high-purity transcriptomes in mouse-human mixing studies, is compatible with multi-omics measurements, and can accurately characterize cell types in human breast tissue when compared to a commercial microfluidic platform. Single-cell transcriptional profiling of mixed phenotype acute leukemia using PIP-seq revealed the emergence of heterogeneity within chemotherapy-resistant cell subsets that were hidden by standard immunophenotyping. PIP-seq is a simple, flexible, and scalable next-generation workflow that extends single-cell sequencing to new applications, including screening, diagnostics, and disease monitoring.

genomics↗

bioRxiv2022

Scalable microbial strain inference in metagenomic data using StrainFacts

While genome databases are nearing a complete catalog of species commonly inhabiting the human gut, their representation of intraspecific diversity is lacking for all but the most abundant and frequently studied taxa. Statistical deconvolution of allele frequencies from shotgun metagenomic data into strain genotypes and relative abundances is a promising approach, but existing methods are limited by computational scalability. Here we introduce StrainFacts, a method for strain deconvolution that enables inference across tens of thousands of metagenomes. We harness a "fuzzy" genotype approximation that makes the underlying graphical model fully differentiable, unlike existing methods. This allows parameter estimates to be optimized with gradient-based methods, speeding up model fitting by two orders of magnitude. A GPU implementation provides additional scalability. Extensive simulations show that StrainFacts can perform strain inference on thousands of metagenomes and has comparable accuracy to more computationally intensive tools. We further validate our strain inferences using single-cell genomic sequencing from a human stool sample. Applying StrainFacts to a collection of more than 10,000 publicly available human stool metagenomes, we quantify patterns of strain diversity, biogeography, and linkage-disequilibrium that agree with and expand on what is known based on existing reference genomes. StrainFacts paves the way for large-scale biogeography and population genetic studies of microbiomes using metagenomic data.

bioinformatics↗

bioRxiv2021

Microbowls with controlled concavity for accurate microscale mass spectrometry

Patterned surfaces can enhance the sensitivity of laser desorption ionization mass spectrometry by segregating and concentrating analytes, but their fabrication can be challenging. Here, we describe a simple method to fabricate substrates patterned with micron-scale wells that yield more accurate and sensitive mass spectrometry measurements compared to flat surfaces. The wells can also concentrate and localize cells and beads for cell-based assays.

bioengineering↗

bioRxiv2021

Multiplexed action-outcome representation by striatal striosome-matrix compartments detected with a novel cost-benefit foraging task

Learning about positive and negative outcomes of actions is crucial for survival and underpinned by conserved circuits including the striatum. How associations between actions and outcomes are formed is not fully understood, particularly when the outcomes have mixed positive and negative features. We developed a novel foraging ( bandit) task requiring mice to maximize rewards while minimizing punishments. By 2-photon Ca++ imaging, we monitored activity of 5831 identified anterodorsal striatal striosomal and matrix neurons. Surprisingly, we found that action-outcome associations for reward and punishment were combinatorially encoded rather than being integrated as overall outcome value. Single neurons could, for one action, encode outcomes of opposing valence. Striosome compartments consistently exhibited stronger representations of reinforcement outcomes than matrix, especially for high reward or punishment prediction errors. These findings demonstrate a remarkable multiplexing of action-outcome contingencies by single identified striatal neurons and suggest that striosomal neurons are differentially important in action-outcome learning.

neuroscience↗

bioRxiv2021

Accurate bulk quantitation of droplet digital PCR

Droplet digital PCR provides superior accuracy in nucleic acid quantitation. The requirement of microfluidics to generate and analyze the emulsions, however, is a barrier to its adoption, particularly in low resource or clinical settings. Here, we report a novel method to prepare ddPCR droplets by vortexing and readout the results by bulk analysis of recovered amplicons. We demonstrate the approach by accurately quantitating SARS-CoV-2 sequences using entirely bulk processing and no microfluidics. Our approach for quantitating reactions should extend to all digital assays that generate amplicons, including digital PCR and LAMP conducted in droplets, microchambers, or nanoliter wells. More broadly, our approach combines important attributes of ddPCR, including enhanced accuracy and robustness to inhibition, with the high-volume sample processing ability of quantitative PCR.

molecular biology↗

bioRxiv2020

Droplet microfluidic sequencing of HIV genomes and integration sites

Sequencing individual HIV-proviruses and their adjacent cellular junctions can elucidate mechanisms of infected cell persistence in vivo. Here, we present a high throughput microfluidic method to sequence entire proviruses in their native integration site context. We used the method to analyze infected cells from people with HIV on suppressive antiretroviral therapy, demonstrating >90% capture and sequencing of paired proviral genomes and integration sites. This method should enable comprehensive genetic analysis of persistent HIV-infected cell reservoirs, providing important insights into the barriers to HIV cure.

genomics↗

bioRxiv2018

High throughput gene expression profiling of yeast colonies with microgel-culture Drop-seq

Yeasts can be engineered into \"living foundries\" for non-natural chemical production by reprogramming their genome using a synthetic biology \"design-build-test\" cycle. While methods for \"design\" and \"build\" are scalable and efficient, \"test\" remains a labor-intensive bottleneck, limiting the effectiveness of the genetic reprogramming results. Here we describe Isogenic Colony Sequencing (ICO-seq), a massively-parallel strategy to assess the gene expression, and thus engineered pathway efficacy, of large numbers of genetically distinct yeast colonies. We use the approach to characterize opaque-white switching in 658 C. albicans colonies. By profiling transcriptomes of 1642 engineered S. cerevisiae strains, we use it to assess gene expression heterogeneity in a protein mutagenesis library. Our approach will accelerate synthetic biology by allowing facile and cost-effective transcriptional profiling of large numbers of genetically distinct yeast strains.

bioengineering↗

bioRxiv2018

Single-cell RNA-seq reveals profibrotic macrophages in lung fibrosis

Myeloid cells localize to peripheral tissues in a wide range of pathologic contexts. However, appreciation of distinct myeloid subtypes has been limited by the signal averaging inherent to bulk sequencing approaches. Here we applied single-cell RNA sequencing (scRNA-seq) to map cellular heterogeneity in lung fibrosis induced by bleomycin injury in mice. We first developed a computational framework that enables unbiased, granular cell-type annotation of scRNA-seq. This approach identified a macrophage subpopulation that was specific to injured lung and notable for high expression of Cx3cr1+ and MHCII genes. We found that these macrophages, which bear a gene expression profile consistent with monocytic origin, progressively acquire alveolar macrophage identity and localize to sites of fibroblast accumulation. Probing their functional role, in vitro studies showed a trophic effect of these cells on fibroblast activation, and ablation of Cx3cr1-expressing cells suppressed fibrosis in vivo. We also found by gene set analysis and immunofluorescence that markers of these macrophages were upregulated in samples from patients with lung fibrosis compared with healthy controls. Taken together, our results uncover a specific pathologic subgroup of macrophages with markers that could enable their therapeutic targeting for fibrosis.

immunology↗