Biology subjects

Gan, D.

Publications and source records attributed to Gan, D..

11 recordsLinked to original sources

bioRxiv2026

Niche-dependent epistasis within the cancer genome shapes metastatic fitness

Epistasis, the genetic interaction between loci, directs non-additive genotype-to-phenotype outcomes. Within a multicellular organism, how the tissue niche (E) shapes gene-by-gene epistasis (GxG) and its phenotypic consequences remains poorly studied. Using tumor lung metastasis as an experimental model, we mapped GxGxE through combinatorial CRISPR perturbation coupled with depletion of defined host myeloid populations in mice. The metastatic niche first selects disseminated tumor cells on GxG epistatic fitness, then drives transcriptome convergence regardless of genotype. Removing anatomically distinct myeloid populations reshaped the metastatic epistasis landscape despite identical tumor genotypes. Gene-by-gene-by-environment epistasis therefore operates at the organismal level, with interaction magnitude and sign set by niche cells. Beyond metastasis, our data suggest that the tissue niche conditions how genetic interactions translate into divergent phenotypes.

genetics↗

bioRxiv2026

Parallel Activation and Interference CRISPR (PAIR) with Sequencing Uncovers DNA Repair Networks Guiding Precision Cell Engineering

The dynamic balance of cellular homeostasis is often maintained by opposing regulatory pathways, yet most genetic screens interrogate them in one direction and therefore miss the bidirectional gene-gene interactions that shape complex phenotypes such as DNA damage response (DDR). Here, we present PAIR (Parallel Activation and Interference CRISPR), a bidirectional perturbation platform that enables simultaneous activation and suppression of distinct genes within the same cell using CRISPR activation (CRISPRa) and Cas13d RNA knockdown. Applying PAIR to the CRISPR/Cas9 induced DSB repair screen, we mapped gene-gene interactions across competing repair branches and identify synergistic perturbations, including NBN activation combined with suppression of end-joining factors, that shift repair outcomes toward homology-directed repair (HDR) and improve the precision of CRISPR-based gene editing. Using coupled PAIR with single-cell transcriptomic, we further demonstrated that NBN activation induces inflammatory and interferon programs, whereas co-suppression of end-joining factors buffers this response, revealing transcriptional states missed by conventional unidirectional perturbations. To translate these findings into non-viral chimeric antigen receptor (CAR) T cell engineering, we developed an mRNA-based strategy for parallel overexpression and knockdown of NBN-anchored DDR effectors in primary T cells, priming the T cells into a transient HDR-favored state that enhances the efficiency of CAR knock-in on the TRAC locus. Together, the PAIR system provides a general framework for studying opposing regulatory networks, uncovering hidden cell states, and guiding cell-state engineering through bidirectional perturbation.

bioengineering↗

bioRxiv2025

Understanding the LLM-based gene embeddings

Large language model (LLM)-derived gene embeddings, generated from brief NCBI gene descriptions, have shown strong performance in recent biological applications, yet the biological information they contain remains unclear. We evaluate these embeddings using a Gene Set Enrichment Analysis (GSEA)-based framework that treats each embedding dimension as a potential carrier of pathway-level information. OpenAIs embeddings recover over 93% of Hallmark and C2 pathways, with pathway signals distributed across many dimensions. Even embeddings generated from gene symbols alone recover more than 64% of pathways, indicating substantial prior biological knowledge embedded in the model. Comparing 11 small language models reveals that domain-specific models perform best with minimal input, but all models approach OpenAI-level coverage when given modest textual context. Collectively, these results show that LLM-derived embeddings encode unexpectedly extensive pathway-level information, supporting their use as lightweight, informative representations for downstream biological analysis.

bioinformatics↗

bioRxiv2025

Genetic dissection of microglia cannibalism reveals an IL10 signaling axis controls microglia lifespan

The development of complex organs, like the brain, demands a robust system for tissue remodeling and cellular debris clearance. In the brain, this function is performed by microglia, which must clear diverse debris substrates, including that caused by cell death. Although the subsequent fate of these phagocytic microglia is a critical regulatory point that impacts whether the brain resolves a debris environment, the genetic mechanisms that control microglia fate after debris clearance remain mostly unknown. To address this, we conducted a large-scale CRISPR screen in zebrafish using a custom-built robotic confocal microscope. We selected candidate genes from a single-cell RNA sequencing dataset of embryonic mouse microglia. This screen identified several modulators of microglial lifespan and cannibalism that are enriched in mouse and zebrafish microglia, including interleukin-10 receptor beta (il10rb), a receptor subunit for the cytokine IL10. Perturbation of il10, il10rb, and downstream signaling molecules JAK/STAT in zebrafish reduced microglial death. Expression analysis in mouse and zebrafish confirmed that microglia express both il10 and il10rb. Given the established role of IL10 in lysosomal remodeling, we hypothesized that it regulates microglial survival through lysosomal acidification. While il10rb perturbation did not alter lysosome number or size, it caused a significant reduction in LysoTracker-positive lysosomes, indicating decreased lysosomal acidification. Inhibiting v-ATPase also reduced microglial death, reinforcing the link between lysosomal pH and cell fate. Our findings reveal a cytokine-regulated mechanism where lysosomal dynamics determine the survival of phagocytic microglia. We propose that a necroptosis-cannibalism process functions as a quality control mechanism for microglial turnover, which is critical for refining neuroimmune cell function in the brain.

neuroscience↗

bioRxiv2025

An oncogenotype-immunophenotype paradigm governing the myeloid landscape in genetically engineered mouse models of prostate cancer

Distinct oncogenotypes sculpt divergent myeloid landscapes in prostate cancer. Using prostate-specific Ptenpc-/-, Ptenpc-/- Trp53pc-/-, Ptenpc-/- Smad4pc-/-, and Ptenpc-/- Trp53pc-/- Smad4pc-/- models, we identify a Smad4-loss-driven neutrophil-enriched subtype (NES) characterized by CXCL5, CXCL2, and CCL20 upregulation, polymorphonuclear myeloid-derived suppressor cell (PMN-MDSC) infiltration, and immune checkpoint resistance. In contrast, Smad4-intact tumors form macrophage-enriched subtypes (MES) responsive to immunotherapy. Mechanistically, Smad4 ablation activates YAP signaling and elevates histone epigenetic regulatory enzymes that enhance histone modification, chromatin accessibility, and transcription of neutrophil-recruiting cytokines. Genetic or pharmacologic inhibition of these enzymes suppresses chemokine expression, reduces neutrophil accumulation, restores CD8 T-cell activity, and limits tumor growth in immunocompetent hosts. Analyses of human prostate cancers support the findings from the murine NES prostate tumors. These findings establish a Smad4-YAP-epigenetic axis linking oncogenotype to immunophenotype and uncover a therapeutic vulnerability within the myeloid-dominant prostate tumor microenvironment.

cancer biology↗

bioRxiv2025

sox17 mutagenesis reveals microglia can be impacted by genetic maternal effects

Embryonic development is precisely shaped by maternal and zygotic factors. These maternal factors exert their influence through maternal effects, a phenomenon where an offsprings phenotype is determined, at least in part, by the mothers environment and genotype. While environmental maternal effects can cause phenotypes that present both early and later in life, genetic maternal effects generally induce phenotypes in the earliest embryonic stages. Here, we reveal a genetic maternal effect that influences the development of cells that arise after early embryogenesis, highlighting that specific cell types can be susceptible to late-onset genetic maternal effects. Using zebrafish to study microglia, the resident immune cells of the brain, we identified a mutation in sry-related HMG box gene-17 (sox17) that exhibits a maternal effect phenotype that presents as a reduction of microglia in the brain and precursors in the yolk sac. We demonstrate that sox17 is expressed in microglia and their yolk sac precursors and is maternally-loaded. We show that sox17 restoration via embryonic injection reverses the maternal effect on microglia and yolk sac cells in sox17 mutants. To identify additional genes interacting with sox17, we nominated genes from scRNA sequencing analysis of mouse embryonic microglia to perform a genetic screen using CRISPR mutagenesis and a custom-built robot that captures confocal images of the zebrafish brain in high-throughput. This screen identified f11r.1, gas6, and mpp1 as modifiers of microglia abundance in the embryonic brain, which we demonstrated are also expressed in zebrafish microglia. Transcriptional and mutant analyses with these new modifiers suggest that sox17 positively regulates mpp1 transcription. These results demonstrate that microglia are susceptible to genetic maternal effects, in addition to their known sensitivity to environmental maternal effects. Our findings reveal a late-onset phenotype associated with the maternal genotype, expanding the recognized impact of genetic maternal effects beyond initial embryo viability and into long-term vigor.

neuroscience↗

bioRxiv2025

Small, Open-Source Text-Embedding Models as Substitutes to OpenAI Models for Gene Analysis

While foundation transformer-based models developed for gene expression data analysis can be costly to train and operate, a recent approach known as GenePT offers a low-cost and highly efficient alternative. GenePT utilizes OpenAIs text-embedding function to encode background information, which is in textual form, about genes. However, the closed-source, online nature of OpenAIs text-embedding service raises concerns regarding data privacy, among other issues. In this paper, we explore the possibility of replacing OpenAIs models with open-source transformer-based text-embedding models. We identified ten models from Hugging Face that are small in size, easy to install, and light in computation. Across all four gene classification tasks we considered, some of these models have outperformed OpenAIs, demonstrating their potential as viable, or even superior, alternatives. Additionally, we find that fine-tuning these models often does not lead to significant improvements in performance.

bioinformatics↗

bioRxiv2023

Overcome Prostate Cancer Resistance to Immune Checkpoint Therapy with Ketogenic Diet-Induced Epigenetic Reprogramming

Advanced prostate cancer (PCa) is overwhelmingly resistant to immune checkpoint blockade (ICB) therapy, representing a formidable clinical challenge. In this study, we developed a syngeneic murine PCa model with acquired ICB resistance. Using this model, synergistic efficacy was achieved by combining anti-PD1 and anti-CTLA4 antibodies with histone deacetylase inhibitor (HDACi) vorinostat, a cyclic ketogenic diet (CKD), or supplementation of ketone body {beta}-hydroxybutyrate (BHB, endogenous HDACi) via 1,3-butanediol-admixed food. CKD and BHB supplementation delayed PCa tumors as monotherapy, and both BHB and adaptive immunity are required for the anti-tumor activity of CKD. Single-cell transcriptomic and proteomic profiling revealed that the HDACi and ketogenesis-enhanced ICB therapy involves cancer-cell-intrinsic (upregulated MHC class I molecules) and extrinsic mechanisms (CD8+ T cell chemoattraction, M1/M2 macrophage rebalancing, monocyte differentiation toward antigen presenting cells, and diminished neutrophils). Overall, these findings underscore the potential of using HDACi and optimized KD to enhance ICB therapy for PCa.

cancer biology↗

bioRxiv2022

The Graphical R2D2 Estimator for the Precision Matrices

Biological networks are important for the analysis of human diseases, which summarize the regulatory interactions and other relationships between different molecules. Understanding and constructing networks for molecules, such as DNA, RNA and proteins, can help elucidate the mechanisms of complex biological systems. The Gaussian Graphical Models (GGMs) are popular tools for the estimation of biological networks. Nonetheless, reconstructing GGMs from high-dimensional datasets is still challenging. Current methods cannot handle the sparsity and high-dimensionality issues arising from datasets very well. Here we developed a new GGM, called the GR2D2 (Graphical R2-induced Dirichlet Decomposition) model, based on the R2D2 priors for linear models. Besides, we provided a data-augmented block Gibbs sampler algorithm. The R code is available at https://github.com/RavenGan/GR2D2. The GR2D2 estimator shows superior performance in estimating the precision matrices compared to existing techniques in various simulation settings. When the true precision matrix is sparse and of high dimension, the GR2D2 provides the estimates with smallest information divergence from the underlying truth. We also compare the GR2D2 estimator to the graphical horseshoe estimator in five cancer RNA-seq gene expression datasets grouped by three cancer types. Our results show that GR2D2 successfully identifies common cancer pathways and cancer-specific pathways for each dataset.

bioinformatics↗

bioRxiv2021

Evolutionary Analysis of Transcriptional Regulation Mediated by Cdx2 in Rodents

Differences in gene expression, which can arise from divergence in cis-regulatory elements or alterations in transcription factors binding specificity, are one of the most important causes of phenotypic diversity during evolution. By protein sequence analysis, we observed high sequence conservation in the DNA binding domain (DBD) of the transcription factor Cdx2 across many vertebrates, whereas three amino acid changes were exclusively found in mouse Cdx2 (mCdx2), suggesting potential positive selection in the mouse lineage. Multi-omics analyses were then carried out to investigate the effects of these changes. Surprisingly, there were no significant functional differences between mCdx2 and its rat homologue (rCdx2), and none of the three amino acid changes had any impact on its function. Finally, we used rat-mouse allodiploid embryonic stem cells (RMES) to study the cis effects of Cdx2-mediated gene regulation between the two rodents. Interestingly, whereas Cdx2 binding is largely divergent between mouse and rat, the transcriptional effect induced by Cdx2 is conserved to a much larger extent. Author summaryOur study 1) represented a first systematic analysis of species-specific adaptation in DNA binding pattern of transcription factor. Although the mouse-specific amino acid changes did not manifest functional impact in our system, several explanations may account for it (See Discussion part for the detail); 2) represented a first study of cis-regulation between two reproductively isolated species by using a novel allodiploid system; 3) demonstrated a higher conservation of transcriptional output than that of DNA binding, suggesting the evolvability/plasticity of the latter; 4) finally provided a rich data resource for Cdx2 mediated regulation, including gene expression, chromatin accessibility and DNA binding etc.

evolutionary biology↗

bioRxiv2020

CIGAR-seq, a CRISPR/Cas-based method for unbiased screening of novel mRNA modification regulators

Cellular RNA is decorated with over 170 types of chemical modifications. Many modifications in mRNA, including m6A and m5C, have been associated with critical cellular functions under physiological and/or pathological conditions. To understand the biological functions of these modifications, it is vital to identify the regulators that modulate the modification rate. However, a high-throughput method for unbiased screening of these regulators is so far lacking. Here, we report such a method combining pooled CRISPR screen and reporters with RNA modification readout, termed CRISPR integrated gRNA and reporter sequencing (CIGAR-seq). Using CIGAR-seq, we discovered NSUN6 as a novel mRNA m5C methyltransferase. Subsequent mRNA bisulfite sequencing in HAP1 cells without or with NSUN6 and/or NSUN2 knockout showed that NSUN6 and NSUN2 worked on non-overlapping subsets of mRNA m5C sites, and together contributed to almost all the m5C modification in mRNA. Finally, using m1A as an example, we demonstrated that CIGAR-seq can be easily adapted for identifying regulators of other mRNA modification.

molecular biology↗