Biomedical subjects

Brian S Plosky

Publications and source records attributed to Brian S Plosky.

8 recordsLinked to original sources

PubMed2026

Tahoe-100M: Mapping drug-induced molecular phenotypes at single-cell resolution.

We present Tahoe-100M, a giga-scale single-cell perturbation atlas comprising 100 million transcriptomes from 50 diverse cancer cell lines treated with 1,100 drug-dose conditions. This parallel profiling of thousands of perturbations at single-cell resolution with minimal batch effects is enabled by the Mosaic platform, which multiplexes genetically distinct cell models into balanced "cell villages." Beyond cataloging transcriptomic shifts, Tahoe-100M systematically quantifies cellular phenotypes, including proliferation, cytotoxicity, lineage-specific vulnerabilities, and cell-cycle changes. It captures population-level transcriptomic heterogeneity, characterizing whether drug responses drive cells toward divergent fates or convergent states. Pathway-based signatures define drug-induced expression programs, classify mechanisms of action, reveal off-target activities, and expose adaptive stress responses associated with resistance. By unifying cellular and molecular readouts, this broadly applicable perturbation atlas advances our ability to model gene regulation, drug response, and network dynamics. Its public release enables the training of AI frameworks to advance predictive models of cell behavior.

Humans↗

PubMed2026

Predicting cellular responses to perturbation across diverse contexts with State.

While machine learning models offer potential for predicting transcriptomic effects of perturbation, they currently struggle to generalize across cellular contexts. Here, we introduce State, a machine learning model that predicts perturbation effects while accounting for cellular heterogeneity within and across experiments. State is trained using single-cell gene expression data to predict perturbation effects across sets of cells. State improved discrimination of effects on large datasets by more than 30% and identified differentially expressed genes across genetic, signaling, and chemical perturbations with significantly improved accuracy compared with baselines. Its cell embeddings trained on observational data from 167 million cells enable the identification of strong perturbations in cellular contexts where no perturbations were observed during training. We further introduce Cell-Eval, a comprehensive evaluation framework that can be used to evaluate future models. Overall, the performance and flexibility of State set the stage for scaling the development of AI models of cell state.

Machine Learning↗

PubMed2006

Controlling the subcellular localization of DNA polymerases iota and eta via interactions with ubiquitin.

Y-family DNA polymerases have spacious active sites that can accommodate a wide variety of geometric distortions. As a consequence, they are considerably more error-prone than high-fidelity replicases. It is hardly surprising, therefore, that the in vivo activity of these polymerases is tightly regulated, so as to minimize their inadvertent access to primer-termini. We report here that one such mechanism employed by human cells relies on a specific and direct interaction between DNA polymerases iota and eta with ubiquitin (Ub). Indeed, we show that both polymerases interact noncovalently with free polyUb chains, as well as mono-ubiquitinated proliferating cell nuclear antigen (Ub-PCNA). Mutants of poliota (P692R) and poleta (H654A) were isolated that are defective in their interactions with polyUb and Ub-PCNA, whilst retaining their ability to interact with unmodified PCNA. Interestingly, the polymerase mutants exhibit significantly lower levels of replication foci in response to DNA damage, thereby highlighting the biological importance of the polymerase-Ub interaction in regulating the access of the TLS polymerases to stalled replication forks in vivo.

DNA-Directed DNA Polymerase↗

PubMed2004

Investigating the role of the little finger domain of Y-family DNA polymerases in low fidelity synthesis and translesion replication.

Dpo4 and Dbh are Y-family polymerases that originate from two closely related strains of Sulfolobaceae. Quite surprisingly, however, the two polymerases exhibit different enzymatic properties in vitro. For example, Dpo4 can replicate past a variety of DNA lesions, yet Dbh does so with a much lower efficiency. When replicating undamaged DNA, Dpo4 is prone to make base pair substitutions, whereas Dbh predominantly makes single-base deletions. Overall, the two proteins are 54% identical, but the greatest divergence is found in their respective little finger (LF) domains, which are only 41% identical. To investigate the role of the LF domain in the fidelity and lesion-bypassing abilities of Y-family polymerases, we have generated chimeras of Dpo4 and Dbh in which their LF domains have been interchanged. Interestingly, by replacing the LF domain of Dbh with that of Dpo4, the enzymatic properties of the chimeric enzyme are more Dpo4-like in that the enzyme is more processive, can bypass an abasic site and a thymine-thymine cyclobutane pyrimidine dimer, and predominantly makes base pair substitutions when replicating undamaged DNA. The converse is true for the Dpo4-LF-Dbh chimera, which is more Dbh-like in its processivity and ability to bypass DNA adducts and generate single-base deletion errors. Our studies indicate that the unique but variable LF domain of Y-family polymerases plays a major role in determining the enzymatic and biological properties of each individual Y-family member.

Amino Acid Sequence↗

PubMed2004

Crystal structure of a benzo[a]pyrene diol epoxide adduct in a ternary complex with a DNA polymerase.

The first occupation-associated cancers to be recognized were the sooty warts (cancers of the scrotum) suffered by chimney sweeps in 18th century England. In the 19th century, high incidences of skin cancers were noted among fuel industry workers. By the early 20th century, malignant skin tumors were produced in laboratory animals by repeatedly painting them with coal tar. The culprit in coal tar that induces cancer was finally isolated in 1933 and determined to be benzo[a]pyrene (BP), a polycyclic aromatic hydrocarbon. A residue of fuel and tobacco combustion and frequently ingested by humans, BP is metabolized in mammals to benzo[a]pyrene diol epoxide (BPDE), which forms covalent DNA adducts and induces tumor growth. In the 70 yr since its isolation, BP has been the most studied carcinogen. Yet, there has been no crystal structure of a BPDE DNA adduct. We report here the crystal structure of a BPDE-adenine adduct base-paired with thymine at a template-primer junction and complexed with the lesion-bypass DNA polymerase Dpo4 and an incoming nucleotide. Two conformations of the BPDE, one intercalated between base pairs and another solvent-exposed in the major groove, are observed. The latter conformation, which can be stabilized by organic solvents that reduce the dielectric constant, seems more favorable for DNA replication by Dpo4. These structures also suggest a mechanism by which mutations are generated during replication of DNA containing BPDE adducts.

7,8-Dihydro-7,8-dihydroxybenzo(a)pyrene 9,10-oxide↗

PubMed2004

Switching from high-fidelity replicases to low-fidelity lesion-bypass polymerases.

Replication of damaged DNA often requires a DNA polymerase in addition to the cell's normal replicase. Recent research has begun to shed light on the switch from a high-fidelity replicative polymerase to a low-fidelity translesion polymerase that occurs at a stalled replication fork. A picture is emerging in which eukaryotic replicative clamps are posttranslationally modified by ubiquitination, SUMOylation or phosphorylation. It is believed that such modifications help to regulate the access of translesion polymerases to the nascent primer terminus.

DNA Damage↗

PubMed2003

129-derived strains of mice are deficient in DNA polymerase iota and have normal immunoglobulin hypermutation.

Recent studies suggest that DNA polymerase eta (poleta) and DNA polymerase iota (poliota) are involved in somatic hypermutation of immunoglobulin variable genes. To test the role of poliota in generating mutations in an animal model, we first characterized the biochemical properties of murine poliota. Like its human counterpart, murine poliota is extremely error-prone when catalyzing synthesis on a variety of DNA templates in vitro. Interestingly, when filling in a 1 base-pair gap, DNA synthesis and subsequent strand displacement was greatest in the presence of both pols iota and eta. Genomic sequence analysis of Poli led to the serendipitous discovery that 129-derived strains of mice have a nonsense codon mutation in exon 2 that abrogates production of poliota. Analysis of hypermutation in variable genes from 129/SvJ (Poli-/-) and C57BL/6J (Poli+/+) mice revealed that the overall frequency and spectrum of mutation were normal in poliota-deficient mice. Thus, either poliota does not participate in hypermutation, or its role is nonessential and can be readily assumed by another low-fidelity polymerase.

Animals↗

PubMed2003

Replication of a cis-syn thymine dimer at atomic resolution.

Ultraviolet light damages DNA by catalysing covalent bond formation between adjacent pyrimidines, generating cis-syn cyclobutane pyrimidine dimers (CPDs) as the most common lesion. CPDs block DNA replication by high-fidelity DNA polymerases, but they can be efficiently bypassed by the Y-family DNA polymerase pol eta. Mutations in POLH encoding pol eta are implicated in nearly 20% of xeroderma pigmentosum, a human disease characterized by extreme sensitivity to sunlight and predisposition to skin cancer. Here we have determined two crystal structures of Dpo4, an archaeal pol eta homologue, complexed with CPD-containing DNA, where the 3' and 5' thymine of the CPD separately serves as a templating base. The 3' thymine of the CPD forms a Watson-Crick base pair with the incoming dideoxyATP, but the 5' thymine forms a Hoogsteen base pair with the dideoxyATP in syn conformation. Dpo4 retains a similar tertiary structure, but each unusual DNA structure is individually fitted into the active site for catalysis. A model of the pol eta-CPD complex built from the crystal structures of Saccharomyces cerevisiae apo-pol eta and the Dpo4-CPD complex suggests unique features that allow pol eta to efficiently bypass CPDs.

Amino Acid Sequence↗