Search arXiv⌕ Search

arXiv subjects

Zijun Gao

Publications and source records attributed to Zijun Gao.

At least 19 recordsLinked to original sources

CellPlex: Benchmarking Single-Cell Spatial Proteomics Prediction from H&E

Spatial proteomics measures tens of proteins in every cell of a tissue section, but its cost confines it to small research cohorts, whereas hematoxylin and eosin (H\&E) slides exist for nearly every patient. Inferring single-cell spatial proteomics from H\&E would bring its analyses to routine pathology, and a growing number of virtual staining models pursue this goal. Each is evaluated on its own cohort, panel and pixel-level metric, and released models transfer poorly to new cohorts, leaving open which approach to build on and which proteins H\&E can reveal. We introduce CellPlex, a benchmark that poses the task at the level of the cell: seven cohorts with same-section H\&E and multiplexed immunofluorescence, from colorectal and breast cancer to lymph-node metastases, bone marrow and glioma, with 717 samples, 44.8 million cells and 149 proteins, processed by one pipeline into per-cell targets. Pathology foundation models (PFMs) are adapted with one cell-level protocol, specialist models are fine-tuned under a matched budget, and all are scored on the same cells. Adapted PFMs are the most accurate models on every cohort, ahead of the specialists even after fine-tuning, but keep less than half of their accuracy on cohorts they were not trained on. Which proteins can be predicted depends on the tissue, not on how many cells were measured for them. By fixing the cells, targets and splits on which predictions are judged, CellPlex sets a common standard for inferring costly molecular measurements from routine H\&E and shows how far H\&E can stand in for spatial proteomics. We will release the data, splits, evaluation code and baselines as a testbed for developing and comparing future pathology foundation models and virtual staining methods.

cs.CV↗

Technical Report: Causal ANOVA Applied to COMPAS

This report applies Causal ANOVA to the COMPAS recidivism prediction dataset to study how demographic variables contribute to variation in two-year recidivism outcomes. Using a directed acyclic graph (DAG) to encode causal assumptions, we estimate counterfactual explainability scores for Race, Sex, and Age Category. We then examine whether the explanatory roles of Race and Sex change after conditioning on age group. The results suggest that Age Category explains the largest share of variation in the full causal model. After conditioning on age, Sex explainability is highest among younger defendants and declines with age, while Race explainability increases with age and is most pronounced among older defendants. As a complementary analysis, we use logistic regression models to interpret the demographic patterns suggested by the Causal ANOVA results. We provide a Python package of Causal ANOVA at https://github.com/devonxiong/causal-anova.

stat.AP↗

Estimation and Inference for Causal Explainability

Understanding how much each variable contributes to an outcome is a central question across disciplines. A causal view of explainability is favorable for its ability in uncovering underlying mechanisms and generalizing to new contexts. Based on a family of causal explainability quantities, we develop methods for their estimation and inference. In particular, we construct a one-step correction estimator using semi-parametric efficiency theory, which explicitly leverages the independence structure of variables to reduce the asymptotic variance. For a null hypothesis on the boundary, i.e., zero explainability, we show its equivalence to Fisher's sharp null, which motivates a randomization-based inference procedure. Finally, we illustrate the empirical efficacy of our approach through simulations as well as an immigration experiment dataset, where we investigate how features and their interactions shape public opinion toward admitting immigrants.

stat.ME↗

Causal Discovery Bench: Can LLMs Autonomously Conduct Effective Experiments?

As large language model (LLM) agents get actively adopted for autonomous research, they must not only reason over existing evidence, but also design informative experiments, allocate resources, analyze sequential observations, and decide when to stop and make conclusions. Yet these capabilities remain poorly measured. We introduce Causal-Discovery-Bench (https://github.com/mshen1019/causal-discovery-bench), a verifiable sandbox with configurable structural equation models and resource constraints to evaluate sequential experimental design and causal discovery without domain-semantic cues. Across 3,802 episodes with 13 models from 5 vendors, we analyze agent behavior along 40 metrics in seven families. Agents' performance exhibits a four-level difficulty hierarchy spanning single-experiment design, sequential exploration of causal relationships, resource-aware experimentation balancing cost and statistical power, and autonomous resolution of conflicting evidence. We identify a substantial performance gap between causal inference from correctly curated evidence and navigating causal discovery through autonomous experimentation: Analyzing agents' action trajectories attributes 94.5% of their mistakes to experimental design rather than to inference, which is further consolidated with ablation studies. Relaxing resource constraints alone does not reliably resolve failures in sequential discovery. In contrast, basic statistical tools substantially improve performance across multiple model families, while advanced tools provide little additional benefit when models lack prerequisite lower-level capabilities.

cs.LG↗

A Gravitationally Lensed Low-Luminosity AGN From The Cosmic Noon

Strong gravitational lensing provides a unique tool of studying intrinsically faint high-redshift active galactic nuclei (AGN). In particular, low-luminosity AGN (LLAGN) can probe black-hole growth in low-mass galaxies that current X-ray surveys cannot otherwise access. We investigate the nature and physical properties of a faint X-ray source triply imaged by the galaxy cluster SDSS J2243$-$0935, and determine whether the source hosts an AGN. We combine archival Hubble Space Telescope imaging and Chandra X-ray observations with the published strong-lensing models of SDSS J2243$-$0935. We constrain the source redshift geometrically, derive magnification maps, and infer the physical properties of the host galaxy. The geometric lensing analysis gives a source redshift of $z=2.14^{+0.12}_{-0.17}$, consistent with the photometric redshift estimate, and a total magnification of $μ=78\pm7$. The source is triply lensed and detected only in the HST F160W band, with an intrinsic magnitude of $m_{\rm AB}=28.2\pm0.2$. The rest-frame optical emission implies a stellar mass of $8.6\lesssim\log(M_\star/M_\odot)\lesssim9.4$. The unobscured star formation rate is constrained to $\mathrm{SFR}_{\rm UV}<0.009\,M_\odot\,\mathrm{yr}^{-1}$, although substantial dust obscuration remains possible. Chandra detects X-ray point-source counterparts to all three HST-detected lensed images. The combined X-ray spectrum contains $100\pm11$ net counts and is consistent with a moderately absorbed power law, with $Γ\simeq 1.48$ and $N_{\rm H}=(5.5\pm4.4)\times10^{22}\,\mathrm{cm}^{-2}$. After correcting for gravitational magnification, we obtain an intrinsic rest-frame $2$--$10$ keV luminosity of $L_{\rm X}=1.2\pm0.3\times10^{42}\,\mathrm{erg\,s^{-1}}$. The X-ray emission is likely dominated by an AGN hosted by a low-mass galaxy at $z\simeq2.14$, making this system a strongly lensed LLAGN at cosmic noon (abridged).

astro-ph.GA↗

Towards Scalable Context-Aware Single-Cell Spatial Transcriptomics Prediction from Histology Images

Predicting gene expression from H&E-stained histology images offers a scalable alternative to costly spatial transcriptomics, yet most existing methods operate at the spot level, where signals from multiple cells are aggregated and critical cellular heterogeneity is obscured. Extending this paradigm to single-cell resolution is non-trivial. Naively applying pathology foundation models faces a scale mismatch: their patch-level representations mix multiple cells, whereas per-cell cropping or resizing distorts morphology and removes local context. Conversely, segmentation-based models without strong pretrained visual encoders often lack the morphological representation capacity needed for accurate molecular prediction and inherit errors from imperfect cell boundary masks. Here, we present CELLO, an efficient end-to-end framework that performs a single pathology foundation model forward pass per image and uses grid sampling to extract location-specific features for all cells simultaneously. We further introduce a distance-decay cross-attention module that refines each cell representation using spatially biased local morphological context. Using 52 public Xenium-H&E pairs from HEST-1k that span 12 organs and approximately 10 million cells, CELLO improves the average predictive accuracy over the evaluated baselines while reducing the mean whole-slide inference time compared to DeepSpot2Cell, a 14.0x speed-up on average that excludes upstream cell segmentation. Our work establishes a scalable foundation for single-cell gene expression prediction from H&E images.

cs.CV↗

Reliable Selection of Heterogeneous Treatment Effect Estimators

We study the problem of selecting the best heterogeneous treatment effect (HTE) estimator from a collection of candidates in settings where the treatment effect is fundamentally unobserved. We cast estimator selection as a multiple testing problem and introduce a ground-truth-free procedure based on a cross-fitted, exponentially weighted test statistic. A key component of our method is a two-way sample splitting scheme that decouples nuisance estimation from weight learning and ensures the stability required for valid inference. Leveraging a stability-based central limit theorem, we establish asymptotic familywise error rate control under mild regularity conditions. Empirically, our procedure provides reliable error control while substantially reducing false selections compared with commonly used methods across ACIC 2016, IHDP, and Twins benchmarks, demonstrating that our method is feasible and powerful even without ground-truth treatment effects.

stat.ML↗

Cubic-Root Gaussian Approximation under Unrestricted Covariance

For Gaussian approximation over high-dimensional rectangles under unrestricted covariance, Chernozhukov et al. (2023b) conjectured that the $n^{-1/4}$ rate, up to logarithmic factors, is near-optimal. We show that, under the coordinatewise subexponential condition with scale $B_n$ and the marginal variance lower bound condition with constant $b$ in Chernozhukov et al. (2023b), the approximation error in dimension $d$ is bounded by \begin{align*} C_b\min\left\{ 1,\, \left(\frac{B_n^2}{n}\right)^{1/3}\{\log(2dn)\}^{7/3} + \frac{B_n}{\sqrt n}\{\log(2dn)\}^{5/2} \right\}. \end{align*} In particular, for bounded $B_n$ and polynomial dimension, the new bound is $n^{-1/3}$ and therefore falsifies the polynomial-dimensional $n^{-1/4}$ near-optimality conjecture. The proof uses a two-stage interpolation and a rank-free matrix-weighted Gaussian surface bound, which may be of independent interest. The initial proof attempt was generated by ChatGPT 5.6 Pro (OpenAI) and subsequently corrected and rewritten by the authors. The machine-checked Lean formalization of the proof can be found at the GitHub repository (https://github.com/WeihanZhang2001/cubic-root-gaussian-approximation-under-unrestricted-covariance).

math.ST↗

Causal Survival Forests with Negative Controls

We study heterogeneous treatment-effect (HTE) estimation in observational survival studies commonly associated with both censored outcomes and unmeasured confounding. We integrate causal survival forests (CSF) with negative controls (NC) from proximal causal inference and introduce Negative Control Causal Survival Forests (NC-CSF), a flexible nonparametric HTE learner for survival analysis. Our approach uses a loss that incorporates proxy variables and Neyman orthogonalization to train the random forest, thereby mitigating bias from unobserved confounding and gaining robustness to nuisance estimation. Through extensive simulations spanning varying levels of confounding, proxy relevance, and censoring mechanisms, we demonstrate that NC-CSF substantially reduces bias and estimation error relative to existing baselines. We further demonstrate the practical utility of our method on various clinical datasets, where it confirms several existing findings and also reveals new interpretable patterns of treatment-effect heterogeneity. To facilitate practical use, we provide an end-to-end Python implementation of NC-CSF that carefully handles implementation details such as nuisance estimation and clipping.

stat.ME↗

Reliable conformal novelty detection at the decision boundary

Novelty detection via conformal $p$-values and BH procedure provides distribution-free global false discovery rate (FDR) control. We present here fundamental limits of this approach by showing that it does not produce reliable detection at the decision boundary. We study boundary false discovery rate (bFDR), the probability that the least extreme reported novelty is in fact a null observation. We first show that the support line (SL) procedure, controlling the bFDR in the continuous independent framework, fails to control the bFDR in the conformal case. We then present several modifications of the SL procedure that restore reliability at the decision boundary, by controlling the bFDR, with specific improvements in situations where many novelties are expected (adaptive procedures) and when the calibration sample is too small with respect to the test sample (subsampled procedures). Numerical experiments with both synthetic and real data support our findings and show the relevance of the new proposed approach.

stat.ME↗

Statistical Inference for Generative Model Comparison

Generative models have achieved remarkable success across a range of applications, yet their evaluation still lacks principled uncertainty quantification. In this paper, we develop a method for comparing how close different generative models are to the underlying distribution of test samples. Particularly, our approach employs the Kullback-Leibler (KL) divergence to measure the distance between a generative model and the unknown test distribution, as KL requires no tuning parameters such as the kernels used by RKHS-based distances, and is the only $f$-divergence that admits a crucial cancellation to enable the uncertainty quantification. Furthermore, we extend our method to comparing conditional generative models and leverage Edgeworth expansions to address limited-data settings. On simulated datasets with known ground truth, we show that our approach realizes effective coverage rates, and has higher power compared to kernel-based methods. When applied to generative models on image and text datasets, our procedure yields conclusions consistent with benchmark metrics but with statistical confidence.

stat.ML↗

Flexible Inference for Winners with Conditional Validity

Researchers often select top-performing options or winners, based on a data-driven criterion, such as treatments, models, or model features and then report effect estimates for the selected winners. Naive post-selection estimates, however, are known to suffer from the winner's curse, producing systematically overoptimistic results. We introduce a flexible conditional inference method that corrects for this overoptimism through an adaptive exponential randomization scheme. Our method achieves selection quality that closely matches that of standard top-k selection, while also yielding shorter confidence intervals than existing approaches. Furthermore, our approach applies broadly to nonparametric settings with asymptotically linear selection statistics, covering wide-ranging applications such as inference for the efficacy of the most promising treatments in clinical trials, the abilities of top-ranked models on leaderboards, and the importance of the most predictive features in a model.

stat.ME↗

Exploring the Alignment of Generation and Understanding in Protein Structure Modeling

Understanding and generation are often treated as two separate paradigms in training deep neural networks, despite the fact that both are trained with closely related objectives such as denoising and masked prediction. While prior studies have shown that generative models often learn suboptimal representations for understanding tasks in vision, it is less understood whether a similar gap exists in the protein domain. In this work, we systematically investigate this question by benchmarking state-of-the-art protein generative models on widely-used protein understanding tasks, and observe that these models exhibit consistently poor performance compared to existing protein encoders. Furthermore, inspired by the Representation Alignment (REPA) framework, we propose to explicitly align generative protein diffusion models with pretrained protein understanding models during training. Experiments on the MotifBench demonstrate that representation alignment significantly improves functional protein generation, boosting the MotifBench score of Protpardelle-1c from 39.2 to 47.1, corresponding to a 20% relative improvement. Our results suggest that representation alignment provides a general and effective mechanism for bridging understanding and generation in protein structure modeling.

cs.CE↗

SF-Cluster: Frustration-Guided MSA Subsampling for Alternative Protein Conformation Recovery

Deep-learning structure predictors are sensitive to their multiple sequence alignment (MSA) input, making MSA subsampling a practical route to recovering alternative conformations. Existing approaches such as AF-Cluster operate in sequence space, providing limited control over which conformational basin is sampled. We introduce SF-Cluster, which subsamples MSAs using patterns of predicted local energetic frustration, a representation largely independent of sequence similarity. Across a benchmark of 48 cases spanning fold-switching, allosteric, oligomerization-coupled, and intrinsically disordered systems, and using an AF-Cluster-style dual-reference RMSD criterion, SF-Cluster improves target-state recovery of the alternative conformation over AF-Cluster across the two-state classes, with the largest improvement observed for allosteric systems (+15.5 percentage points). The selected MSAs transfer to an architecturally distinct predictor, indicating that the conformational signal resides in MSA composition. Mechanistically, matched-depth controls show that this recovery advantage is largely explained by the effective depth of the selected subsets, which frustration-pattern selection reliably reaches. At the same time, highly frustrated residues are enriched at sites supported by deep mutational scanning and NMR two-state exchange, and frustration covariation is enriched at state-switching contacts while remaining distinct from coevolutionary coupling. Together, these results identify frustration patterns as a transferable representation for conformational prediction and position MSA subsampling as a representation-guided reweighting problem.

q-bio.BM↗

Partial Identification under High-Dimensional Potential Outcomes and Confounders via Optimal Transport

Partial identification provides informative causal guarantees when point identification is impossible, but existing approaches based on optimal transport (OT) become computationally and statistically intractable in high-dimensional settings. This limitation is particularly severe when both potential outcomes and confounders are high-dimensional, where classical OT-based bounds suffer from the curse of dimensionality and unfavorable convergence rates. To address this challenge, we propose a novel estimator that decomposes the transport problem into a low-dimensional signal subspace and a high-dimensional residual subspace. Unlike existing projection-based methods that discard residual information, we recover the residual transport energy using the Sliced Wasserstein distance, which is computationally efficient and robust to high dimensions. We establish interpretable conditions controlling the approximation gap based on residual structure and provide a data-driven rule for signal dimension selection. Empirical results show that our estimator consistently outperforms projection-only baselines by recovering lost transport energy, yielding more informative causal bounds while remaining computationally tractable in high dimensions.

stat.ME↗

Selective Randomization Inference for Adaptive Experiments

Adaptive experiments use preliminary analyses of the data to inform further course of action and are commonly used in many disciplines including medical and social sciences. Because the null hypothesis and experimental design are data-dependent, it has long been recognized that statistical inference for adaptive experiments is not straightforward. Most existing methods only apply to specific adaptive designs and rely on strong assumptions. In this work, we propose selective randomization inference as a general framework for analysing adaptive experiments. In a nutshell, our approach applies conditional post-selection inference to randomization tests. By using directed acyclic graphs to describe the data generating process, we derive a selective randomization p-value that controls the selective type-I error. As inference only relies on the randomness in the treatment assignment, no modelling assumptions or independent and identically distributed data are needed. We elaborate on conditions that render the proposed p-value computable and provide rejection sampling and MCMC algorithms to find a Monte Carlo approximation. Moreover, this article shows how to estimate and construct confidence intervals for a homogeneous treatment effect. Lastly, we demonstrate our method and compare it with other randomization tests using synthetic and real-world data.

stat.ME↗

GeoCycler: Reward-Aligned 3D Diffusion for Constraint-Conditioned Cyclic Peptide Design

Cyclic peptides are attractive therapeutic modalities because their closed-ring topology can improve stability and target specificity. However, de novo cyclic peptide design remains challenging for diffusion generators, as macrocyclization requires satisfying sparse, non-smooth, and compositional geometric constraints. Existing constraint-conditioned methods largely rely on inference-time guidance, which can steer samples toward desired closures but does not directly change the learned generative distribution. We propose GeoCycler, a reward-weighted diffusion alignment framework for training conditional latent diffusion models toward macrocyclization feasibility. GeoCycler introduces a type-gated stair reward that activates distance-based shaping only when prerequisite residue or linker types are satisfied, providing dense geometric feedback while avoiding misleading signals from chemically incompatible anchors. Together with positive-only reward weighting and replay-based stabilization, GeoCycler aligns a single generator across multiple cyclization topologies. On the LNR benchmark, GeoCycler improves pass@5 closure success over strong guidance-based baselines across stapled, head-to-tail, disulfide, and bicyclic settings. In particular, it improves head-to-tail success by 20.8 percentage points over CP-Composer while maintaining comparable amino-acid and backbone-dihedral statistics. These results suggest that training-time alignment to sparse geometric constraints is a promising alternative to relying solely on post hoc sampling-time correction for cyclic peptide generation.

cs.CE↗

Simultaneous false discovery rate control in location families

When testing a number of statistical hypotheses using data from location families, it is often useful to control the false discovery rate (FDR) not just for hypotheses of the null values but also of other parameter values that are deemed practically insignificant. Here we consider FDR as a curve indexed by the location parameter and suggest a simple generalization of the Benjamini-Hochberg procedure that controls the FDR curve below any user-specified level. As a corollary of our main result, we show that the standard Benjamini-Hochberg procedure -- designed to control the FDR at the null -- also provides simultaneous control of the whole FDR curve for free. We further demonstrate the implications of our results and some practical considerations with a numerical example.

stat.ME↗