Biology subjects

Kumar, R. J.

Publications and source records attributed to Kumar, R. J..

5 recordsLinked to original sources

bioRxiv2026

Generative Language Modeling for Antibody CDR Grafting and Alignment-driven De Novo Design

Antibodies recognise their targets through hypervariable complementarity-determining regions (CDRs), which are interleaved with conserved frameworks in sequence space, making de novo CDR design an infilling problem. Autoregressive models generate residues left-to-right, which precludes full framework context during CDR generation and conflates framework and CDR likelihoods, leaving no natural prompt-response interface for feedback to steer generation. We present GenCDR, a family of LLaMa-based autoregressive language models that read all frameworks as a conditioning prompt and generate all CDRs jointly as a variable-length response, making CDR likelihoods a clean, separable target for reward attribution. The family comprises IgGenCDR, p-IgGenCDR, and NanoGenCDR, trained on unpaired, paired, and nanobody chains, respectively. GenCDR achieves the highest CDR recovery among autoregressive models and produces natural, diverse, human-like CDRs whose likelihoods correlate with fitness and developability assays. The prompt-response boundary also enables principled alignment: reward signals for binding affinity, expression, or developability can be composed to steer CDR generation. Over four rounds of alignment against antibody-antigen co-folding and developability objectives, we find that NanoGenCDR, which uses no explicit antigen encoding, can reach in silico structural interface metrics competitive with those of a structure-conditioned diffusion pipeline at roughly half the sampling budget, with more natural, developable designs. The same interface can be extended to integrate experimental feedback, opening a path to closed-loop antibody de novo design.

synthetic biology↗

bioRxiv2026

Ploidy shapes gemcitabine response through altered potency and delayed cell death

Aberrant tumor ploidy is a near-universal hallmark of cancer and increasingly recognized as a determinant of therapeutic response, but the mechanisms by which ploidy shapes sensitivity to specific cytotoxic agents remain unclear. Here, we investigated the relationship between ploidy and therapeutic response using pharmacogenomic reanalysis, isogenic cancer cell systems, live-cell imaging, intracellular pharmacokinetic/ pharmacodynamic (PK/PD) measurements, and mathematical modeling. Across public pharmacogenomic datasets, gemcitabine emerged as a low-ploidy-selective cytotoxic agent. In matched isogenic low- and high-ploidy cell systems, higher-ploidy cells were consistently less sensitive to gemcitabine across multiple lineages. Live-cell imaging and PK/PD measurements in near-diploid and near-tetraploid SUM-159 cells showed that both states formed intracellular dFdCTP, active form of gemcitabine; but, high-ploidy cells exhibited weaker and slower treatment responses, with delayed accumulation of cell death. To quantify these differences, we developed a delay-aware live/dead model driven by intracellular dFdCTP exposure. The model identified both reduced effective gemcitabine potency and a substantially longer delay from intracellular drug action to observed death in high-ploidy cells (17.5 hours in near-diploid cells versus 42.5 hours in near-tetraploid cells). Interpreting these fitted quantities alongside checkpoint signaling and metabolomic profiling suggests that high-ploidy cells convert intracellular gemcitabine exposure less efficiently into replication-stress signaling, nucleotide-metabolic disruption, and cytotoxic commitment. Together, these results establish ploidy as a determinant of both the magnitude and timing of gemcitabine response and provide a quantitative framework for linking intracellular drug exposure to delayed cytotoxic outcomes across ploidy states.

cancer biology↗

bioRxiv2022

Mre11 liberates cGAS from nucleosome sequestration during tumorigenesis

Oncogene-induced replication stress generates endogenous DNA damage that activates cGAS/STING-mediated innate immune signaling and tumor suppression1-3. However, the mechanism for cGAS activation by endogenous DNA damage remains enigmatic, particularly given the constitutive inhibition of cGAS by high-affinity histone acidic patch (AP) binding4-10. Here we report an in vivo CRISPR screen that identified the DNA double strand break sensor Mre11 as a suppressor of mammary tumorigenesis induced by Myc overexpression and p53 deficiency. Mre11 antagonizes Myc-induced proliferation through cGAS/STING activation. Direct binding of the Mre11-Rad50-Nbn (MRN) complex to nucleosomes displaces cGAS from AP sequestration, which is required for DNA damage-induced cGAS mobilization and activation by cytosolic DNA. Mre11 is thereby essential for cGAS activation in response to oncogenic stress, cytosolic DNA transfection, and ionizing radiation. Furthermore, we show Mre11-dependent cGAS activation suppresses Myc-induced proliferation through ZBP1/RIPK3/MLKL-mediated necroptosis. In human triple-negative breast cancer, ZBP1 downregulation correlates with increased genome instability, decreased immune infiltration, and poor patient prognosis. These findings establish Mre11 as a critical link between DNA damage and cGAS activation that regulates tumorigenesis through ZBP1-dependent necroptosis. One-sentence summaryMre11 is required for cGAS activation during oncogenic stress and promotes ZBP1-dependent necroptosis.

cell biology↗

bioRxiv2020

Hyperactive end joining repair mediates resistance to DNA damaging therapy in p53-deficient cells

TP53 mutations in cancer are associated with poor patient outcomes and resistance to DNA damaging therapies1-3. However, the mechanisms underlying treatment resistance in p53-deficient cells remain poorly characterized. Here, we show that p53-deficient cells exhibit hyperactive repair of therapy-induced DNA double strand breaks (DSBs), which is suppressed by inhibition of DNA-dependent protein kinase (DNA-PK). Single-cell analyses of DSB repair kinetics and cell cycle state transitions reveal an essential role for DNA-PK in suppressing S phase DNA damage and mitotic catastrophe in p53-deficient cells. Yet, a subset of p53-deficient cells exhibit intrinsic resistance to therapeutic DSBs due to a repair pathway that is not sensitive to DNA-PK inhibition. We show that p53 deficiency induces overexpression of DNA Polymerase Theta (Pol {theta}), which mediates an alternative end-joining repair pathway that becomes hyperactivated by DNA-PK inhibition4. Combined inhibition of DNA-PK and Pol {theta} restores therapeutic DNA damage sensitivity in p53-deficient cells. Thus, our study identifies two targetable DSB end joining pathways that can be suppressed as a strategy to overcome resistance to DNA-damaging therapies in p53-deficient cancers.

cell biology↗

bioRxiv2018

Evidence that the cell cycle is a series of uncoupled, memoryless phases

The cell cycle is canonically described as a series of 4 phases: G1 (gap phase 1), S (DNA synthesis), G2 (gap phase 2), and M (mitosis). Various models have been proposed to describe the durations of each phase, including a two-state model with fixed S-G2-M duration and random G1 duration1,2; a \"stretched\" model in which phase durations are proportional3; and an inheritance model in which sister cells show correlated phase durations2,4. A fundamental challenge is to understand the quantitative laws that govern cell-cycle progression and to reconcile the evidence supporting these different models. Here, we used time-lapse fluorescence microscopy to quantify the durations of G1, S, G2, and M phases for thousands of individual cells from three human cell lines. We found no evidence of correlation between any pair of phase durations. Instead, each phase followed an Erlang distribution with a characteristic rate and number of steps. These observations suggest that each cell cycle phase is memoryless with respect to previous phase durations. We challenged this model by perturbing the durations of specific phases through oncogene activation, inhibition of DNA synthesis, reduced temperature, and DNA damage. Phase durations remained uncoupled in individual cells despite large changes in durations in cell populations. To explain this behavior, we propose a mathematical model in which the independence of cell-cycle phase durations arises from a large number of molecular factors that each exerts a minor influence on the rate of cell-cycle progression. The model predicts that it is possible to force correlations between phases by making large perturbations to a single factor that contributes to more than one phase duration, which we confirmed experimentally by inhibiting cyclin-dependent kinase 2 (CDK2). We further report that phases can show coupling under certain dysfunctional states such as in a transformed cell line with defective cell cycle checkpoints. This quantitative model of cell cycle progression explains the paradoxical observation that phase durations are both inherited and independent and suggests how cell cycle progression may be altered in disease states.

systems biology↗