Adrian T. Grzybowski
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Adrian T. Grzybowski has a total of 155 co-authors. Here are the top co-authors by number of shared publications:
- Alexander Jackson Ruthenburg
University of Chicago
18 shared publications Christopher Cafariello
9 shared publications- Christopher J. Porter
Ottawa Bioinformatics Core Facility, Ottawa Hospital Research Institute, Ottawa, Canada
9 shared publications Hani Jrade
9 shared publicationsHannah L. Battaion
9 shared publicationsJanet L. Manias Rothberg
9 shared publications
Latest Publications
- Open Access
TRAILBLAZER: generative multicellular perturbation model of biology
Single-cell foundation models are reshaping biology by learning transferable representations of cellular state from millions of profiles. These models support annotation, denoising, cross-modal mapping and, increasingly, prediction of responses to genetic or pharmacological perturbations. Despite this progress, most approaches treat cells as independent observations and ignore the multicellular context that governs tissue behavior. Models trained on aggregated datasets often fail to generalize to new donors, laboratories or interventions, in part because their latent spaces lack structure for composition and extrapolation. As a result, strong reconstruction performance does not guarantee accurate forecasting of system-level responses. The general problem addressed here is how to construct a scalable model that predicts multicellular, patient-level responses to interventions while preserving single-cell resolution and enabling generalization beyond observed conditions. Here we show that TRAILBLAZER, a multicellular transformer encoder coupled to an explicitly shaped hyperspherical latent space and a count-aware generative decoder, enables accurate zero-shot prediction of perturbation responses and ranking of candidate immunomodulators at patient resolution. In contrast to prior single-cell or pseudo-bulk approaches, TRAILBLAZER models tissues as coordinated systems using latent tokens that summarize and redistribute global context while maintaining near-linear scaling with group size. By organizing latent geometry around shared healthy references and calibrated mechanistic directions, the model renders vector arithmetic biologically meaningful and supports extrapolation to unseen agents. Together, these results establish a practical framework for mechanism-aware simulation of multicellular responses and suggest a path toward predictive foundation models for therapeutic discovery.
Open AccessTRAILBLAZER: generative multicellular perturbation model of biology
Single-cell foundation models are reshaping biology by learning transferable representations of cellular state from millions of profiles. These models support annotation, denoising, cross-modal mapping and, increasingly, prediction of responses to genetic or pharmacological perturbations. Despite this progress, most approaches treat cells as independent observations and ignore the multicellular context that governs tissue behavior. Models trained on aggregated datasets often fail to generalize to new donors, laboratories or interventions, in part because their latent spaces lack structure for composition and extrapolation. As a result, strong reconstruction performance does not guarantee accurate forecasting of system-level responses. The general problem addressed here is how to construct a scalable model that predicts multicellular, patient-level responses to interventions while preserving single-cell resolution and enabling generalization beyond observed conditions. Here we show that TRAILBLAZER, a multicellular transformer encoder coupled to an explicitly shaped hyperspherical latent space and a count-aware generative decoder, enables accurate zero-shot prediction of perturbation responses and ranking of candidate immunomodulators at patient resolution. In contrast to prior single-cell or pseudo-bulk approaches, TRAILBLAZER models tissues as coordinated systems using latent tokens that summarize and redistribute global context while maintaining near-linear scaling with group size. By organizing latent geometry around shared healthy references and calibrated mechanistic directions, the model renders vector arithmetic biologically meaningful and supports extrapolation to unseen agents. Together, these results establish a practical framework for mechanism-aware simulation of multicellular responses and suggest a path toward predictive foundation models for therapeutic discovery.
- 6 March 2024
- Published by Springer Nature in Nature Structural & Molecular Biology
- Vol. 31 (7) , 1039-1049
Guiding the HBO1 complex function through the JADE subunit
JADE is a core subunit of the HBO1 acetyltransferase complex that regulates developmental and epigenetic programs and promotes gene transcription. Here we describe the mechanism by which JADE facilitates recruitment of the HBO1 complex to chromatin and mediates its enzymatic activity. Structural, genomic and complex assembly in vivo studies show that the PZP (PHD1–zinc-knuckle–PHD2) domain of JADE engages the nucleosome through binding to histone H3 and DNA and is necessary for the association with chromatin targets. Recognition of unmethylated H3K4 by PZP directs enzymatic activity of the complex toward histone H4 acetylation, whereas H3K4 hypermethylation alters histone substrate selectivity. We demonstrate that PZP contributes to leukemogenesis, augmenting transforming activity of the NUP98–JADE2 fusion. Our findings highlight biological consequences and the impact of the intact JADE subunit on genomic recruitment, enzymatic function and pathological activity of the HBO1 complex.
- 6 March 2024
- Published by Springer Nature in Nature Structural & Molecular Biology
- Vol. 31 (7) , 1039-1049
Guiding the HBO1 complex function through the JADE subunit
JADE is a core subunit of the HBO1 acetyltransferase complex that regulates developmental and epigenetic programs and promotes gene transcription. Here we describe the mechanism by which JADE facilitates recruitment of the HBO1 complex to chromatin and mediates its enzymatic activity. Structural, genomic and complex assembly in vivo studies show that the PZP (PHD1–zinc-knuckle–PHD2) domain of JADE engages the nucleosome through binding to histone H3 and DNA and is necessary for the association with chromatin targets. Recognition of unmethylated H3K4 by PZP directs enzymatic activity of the complex toward histone H4 acetylation, whereas H3K4 hypermethylation alters histone substrate selectivity. We demonstrate that PZP contributes to leukemogenesis, augmenting transforming activity of the NUP98–JADE2 fusion. Our findings highlight biological consequences and the impact of the intact JADE subunit on genomic recruitment, enzymatic function and pathological activity of the HBO1 complex.
- Open Access
- 3 April 2023
- Published by American Association for Cancer Research (AACR)
Supplementary Table 2: H2B from A Mutation in Histone H2B Represents a New Class of Oncogenic Driver
- 3 April 2023
- Published by American Association for Cancer Research (AACR)
Open AccessSupplementary Table 2: H2B from A Mutation in Histone H2B Represents a New Class of Oncogenic Driver
- Open Access
- 3 April 2023
- Published by American Association for Cancer Research (AACR)
Supplementary Table 3: H3 from A Mutation in Histone H2B Represents a New Class of Oncogenic Driver
- 3 April 2023
- Published by American Association for Cancer Research (AACR)
Open AccessSupplementary Table 3: H3 from A Mutation in Histone H2B Represents a New Class of Oncogenic Driver
- Open Access
- 3 April 2023
- Published by American Association for Cancer Research (AACR)
Supplemental Text and Figures from A Mutation in Histone H2B Represents a New Class of Oncogenic Driver
- 3 April 2023
- Published by American Association for Cancer Research (AACR)
Open AccessSupplemental Text and Figures from A Mutation in Histone H2B Represents a New Class of Oncogenic Driver