Ahn, N. | Search bioRxiv

bioRxiv2026

CytoGate-Bench: an LLM benchmark for cross-panel cell gating in cytometry

In cytometry, the workhorse single-cell technology of clinical immunology, every study defines its own antibody panel and cell-type vocabulary, so a classifier trained on one cannot annotate the next. Immunologists instead annotate by manual gating, splitting one parent population at a time on a two-marker plot, down an expert-defined hierarchy. We introduce CytoGate-Bench, a benchmark that reformulates this per-step procedure as a zero-shot, panel-agnostic task for large language models. It comprises 23,646 expert-annotated instances re-curated from 11 public flow- and mass-cytometry cohorts spanning eight marker panels. Across six open- and closed-weight backbones, the strongest formulation draws one rectangular gate per candidate and falls within the range of trained, panel-specialized baselines. It degrades less under distribution shift. Walking the hierarchy stepwise outperforms predicting every cell type at once. Ablations trace the signal to the data distribution shape and curated marker priors. However, adding vision or a self-verification loop systematically tightens gates.

bioinformatics↗

bioRxiv2022

Recognition of HIV-1 capsid licenses innate immune response to viral infection

Cyclic GMP-AMP synthase (cGAS) is a primary sensor of aberrant DNA that governs an innate immune signaling cascade, leading to the induction of the type-I interferon response. We have previously identified polyglutamine binding protein 1, PQBP1, as an adaptor molecule required for cGAS-mediated innate immune response of lentiviruses, including the human immunodeficiency virus 1 (HIV-1), but dispensable for the recognition of DNA viruses. HIV-1- encoded DNA is synthesized as a single copy from its RNA genome, and is subsequently integrated into the host chromatin. HIV-1 then produces progeny through amplification and packaging of its RNA genome, thus, in contrast to DNA viruses, HIV-1 DNA is both transient and of low abundance. However, the molecular basis for the detection and verification of this low abundance HIV-1 DNA pathogen-associated molecular pattern (PAMP) is not understood. Here, we elucidate a two-factor authentication strategy that is employed by the innate immune surveillance machinery to selectively respond to the low concentration of PAMP, while discerning these species from extranuclear DNA molecules. We find that, upon HIV-1 infection, PQBP1 decorates intact viral capsid, which serves as a primary verification step for the viral nucleic acid cargo. As the reverse transcription and capsid disassembly initiate, cGAS protein is then recruited to the capsid in a PQBP1-dependent manner, enabling cGAS molecules to be co-positioned at the site of PAMP generation. Thus, these data indicate that PQBP1 recognition of the HIV-1 capsid sanctions a robust cGAS-dependent response to a limited abundance and short-lived DNA PAMP. Critically, this illuminates a molecular strategy wherein the modular recruitment of co-factors to germline encoded pattern recognition receptors (PRRs) serves to enhance repertoire of pathogens that can be sensed by the innate immune surveillance machinery.

molecular biology↗