Biology subjects

Ho, A. J.

Publications and source records attributed to Ho, A. J..

4 recordsLinked to original sources

bioRxiv2026

GeneBench-Pro: Evaluating Multistage Statistical Reasoning\\in Genomics, Quantitative Biology, and Translational Biomedicine

We introduce GeneBench-Pro, an expanded and improved version of GeneBench that comprises harder problems across a wider breadth of domains. GeneBench-Pro is a benchmark for AI agents performing realistic multi-stage scientific analyses in genomics, quantitative biology, and translational biomedicine which seeks to capture the complexity of real-world problems that computational life scientists face when tasked with producing a conclusion upon which a downstream scientific or translational decision is contingent. The benchmark comprises 129 evaluations targeting quantities of direct practical relevance across 10 primary domains and 21 terminal subdomains, with a genomics-centered core. Similarly to GeneBench, each problem provides the agent with brief context, a target estimand, and minimal guidance otherwise; the agent must then navigate multiple dependent decision points; i.e., substantive inferential forks where a plausible wrong choice changes the downstream analysis, to identify and execute the correct analysis workflow and arrive at the correct answer. Relative to GeneBench, GeneBench-Pro adds 29 new problems, drops three, and introduces significantly redesigned versions of 54 of the remaining 100 overlapping problems. 82 of the 129 problems were reviewed by external domain experts, whose findings led to prompt/data modifications and redesign of those problems whose targets were not sufficiently identifiable. Ten externally reviewed problems are released publicly, 50 held-out problems were provided to Artificial Analysis for independent third-party model benchmarking, and the remainder are retained as an internal holdout. In evaluations over the full 129-problem suite, GPT-5.6 Sol reaches an eval-level pass rate of 28.7% at the max reasoning level, and GPT-5.6 Sol Pro reaches 31.5% in separately reported GPT Pro runs. GPT-5.5 reaches 12.0%, GPT-5.4 reaches 8.9%, and the strongest non-GPT baseline, Claude Opus 4.8, reaches 16.0%. As with GeneBench, models often complete substantial portions of the workflow but exhibit a consistent gap between noticing and acting by identifying local diagnostic signals but failing to propagate the implications to the corresponding analysis decision. As a result, models often select wrong estimators or persist on initially plausible but incorrect analysis paths. GeneBench-Pro therefore measures an emerging capability of long-horizon biological reasoning that remains unreliable. O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=46 SRC="FIGDIR/small/735386v4_ufig1.gif" ALT="Figure 1"> View larger version (13K): org.highwire.dtl.DTLVardef@191f09org.highwire.dtl.DTLVardef@144943borg.highwire.dtl.DTLVardef@15fba72org.highwire.dtl.DTLVardef@1c9c65c_HPS_FORMAT_FIGEXP M_FIG C_FIG

bioinformatics↗

bioRxiv2025

Joint single-cell profiling of CRISPR-Cas9 edits and transcriptomes reveals widespread off-target events and their effects on gene expression

A longstanding barrier in genome engineering with CRISPR-Cas9 has been the inability to measure editing outcomes and their functional effects at single-cell resolution. Here we present "Superb-seq", which captures single-cell edits and measures their effects on the transcriptome by integrating Cas9 editing, T7 transcription, and single-cell RNA sequencing. We performed Superb-seq on over 34,000 cells from three cell lines. Using seven guide RNAs to target four chromatin remodeler genes in 10,000 K562 cells, Superb-seq detected 43 total edit sites, including 36 off-target sites, and 5761 edited cells (up to 5 edits per cell). Superb-seq improved estimation of gene perturbation effects compared to Perturb-seq and its edit detection sensitivity was comparable to bulk off-target detection methods. We identified 19 off-target edits associated with differential gene expression, nine of which were cell type specific. In summary, Superb-seq illuminates the consequences of Cas9 genome editing by connecting detected single-cell edits with changes in gene expression.

genomics↗

bioRxiv2022

Cocaine addiction-like behaviors are associated with long-term changes in gene regulation, energy metabolism, and GABAergic inhibition within the amygdala

The amygdala processes positive and negative valence and contributes to the development of addiction, but the underlying cell type-specific gene regulatory programs are unknown. We generated an atlas of single nucleus gene expression and chromatin accessibility in the amygdala of outbred rats with low and high cocaine addiction-like behaviors following prolonged abstinence. Between rats with different addiction indexes, we identified thousands of cell type-specific differentially expressed genes enriched for energy metabolism-related pathways that are known to affect synaptic transmission and action potentials. Rats with high addiction-like behaviors showed enhanced GABAergic transmission in the amygdala, which, along with relapse-like behaviors, were reversed by inhibition of Glyoxalase 1, which metabolizes the GABAA receptor agonist methylglyoxal. Finally, we identified thousands of cell type-specific chromatin accessible sites and transcription factor (TF) motifs where accessibility was associated with addiction index, most notably at motifs for pioneer TFs in the Fox, Sox, helix-loop-helix, and AP1 families.

genomics↗

bioRxiv2022

Deletion mapping of regulatory elements for GATA3 reveals a distal T helper 2 cell enhancer involved in allergic diseases

The GATA3 gene is essential for T cell differentiation and is surrounded by risk variants for immune traits. Interpretation of these variants is challenging because the regulatory landscape of GATA3 is complex with dozens of potential enhancers spread across a large topological associating domain (TAD) and gene expression quantitative trait locus (eQTL) studies provide limited evidence for variant function. Here, we perform a tiling deletion screen in Jurkat T cells to identify 23 candidate regulatory elements. Using small deletions in primary T helper 2 (Th2) cells, we validate the function of five of these elements, two of which contain risk variants for asthma and allergic diseases. We fine-map genome-wide association study (GWAS) signals in a distal regulatory element, 1 Mb downstream, to identify 14 candidate causal variants. Small deletions spanning candidate rs725861 decrease GATA3 expression in Th2 cells suggesting a causal mechanism for this variant in allergic diseases. Our study demonstrates the power of integrating GWAS signals with deletion mapping and identifies critical regulatory sequences for GATA3.

genomics↗