Biology subjects

Chitra, U.

Publications and source records attributed to Chitra, U..

11 recordsLinked to original sources

bioRxiv2026

Evaluating the ability of spatial transcriptomics foundation models to learn multi-scale spatial variation

Spatial gene expression results from the superposition of multiple sources of variation in gene expression across different spatial scales, including local microenvironment-associated variation and global spatial gradients. Spatial foundation models (SFMs) are large-scale machine learning models trained on cohorts of spatial transcriptomics (ST) data that, in principle, learn the different sources of spatial variation in gene expression. However, the embeddings learned by SFMs are difficult to interpret, and it remains unclear whether they fully capture such spatial variation. Here, we develop SAFFRON, a sparse autoencoder (SAE)-based framework for interpreting and evaluating SFMs. SAFFRON uses a Matryoshka SAE to decompose dense SFM embeddings into sparse, human-interpretable features and evaluates whether these features correlate with known sources of spatial variation. Using SAFFRON, we systematically benchmark the ability of several recent SFMs to identify local and global spatial variation in gene expression. We find that one SFM, Novae, learns global spatial gradients more accurately than naive, non-foundation model baselines, and that these gradients are concentrated in a small subset of sparse and human-interpretable SAE features revealed by SAFFRON. On the other hand, no SFM learns local microenvironment-associated patterns more accurately than such baselines. Our findings suggest that current SFMs do not systematically learn multi-scale spatial variation in gene expression. CodeSAFFRON is available at https://github.com/chitra-lab/SAFFRON.

bioinformatics↗

bioRxiv2025

Mapping spatial gradients in spatial transcriptomics data with score matching

Spatial transcriptomics (ST) technologies measure gene expression at thousands of locations within a two-dimensional tissue slice, enabling the study of spatial gene expression patterns. Spatial variation in gene expression is characterized by spatial gradients, or the collection of vector fields describing the direction and magnitude in which the expression of each gene increases. However, the few existing methods that learn spatial gradients from ST data either make restrictive and unrealistic assumptions on the structure of the spatial gradients or do not accurately model discrete transcript locations/counts. We introduce SLOPER (for Score-based Learning Of Poisson-modeled Expression Rates), a generative model for learning spatial gradients (vector fields) from ST data. SLOPER models the spatial distribution of mRNA transcripts with an inhomogeneous Poisson point process (IPPP) and uses score matching to learn spatial gradients for each gene. SLOPER utilizes the learned spatial gradients in a novel diffusion-based sampling approach to enhance the spatial coherence and specificity of the observed gene expression measurements. We demonstrate that the spatial gradients and enhanced gene expression representations learned by SLOPER leads to more accurate identification of tissue organization, spatially variable gene modules, and continuous axes of spatial variation (isodepth) compared to existing methods. Software availabilitySLOPER is available at https://github.com/chitra-lab/SLOPER.

bioinformatics↗

bioRxiv2025

Spatial metabolic gradients in the liver and small intestine

Mammalian organs are composed of cells, whose properties differ depending upon their spatial location. Gene expression in liver varies between periportal and pericentral hepatocytes1-3, and in intestine from crypts to villus tips4,5. A key element of tissue spatial organization is likely metabolic, but direct assessments of spatial metabolism remain limited. Here we map spatial metabolic gradients in murine liver and intestine. We developed an integrated experimental-computational workflow using MALDI imaging mass spectrometry, isotope tracing, and deep-learning artificial intelligence. Most measured metabolites (> 95%) showed significant spatial concentration gradients in liver lobules and intestinal villi. In the liver, tricarboxylic acid (TCA)-cycle metabolites and their labeling from both glutamine and lactate localized periportally. Energy-stress metabolites including adenosine monophosphate (AMP) also localized periportally, consistent with high periportal energy demand. In intestine, the TCA intermediates malate (tip) and citrate (crypt) showed opposite spatial patterns, which aligned with higher glutamine catabolism in tips and lactate oxidation in crypts based on isotope tracing. Finally, we mapped the fate of the obesogenic dietary sugar fructose. In the intestine, oral fructose was catabolized faster in the villus bottom than the tips. In the liver, fructose-derived carbon accumulated pericentrally as fructose-1-phosphate and triggered pericentral adenosine triphosphate (ATP) depletion. Thus, we both provide foundational knowledge regarding intestine and liver metabolic organization and identify fructose-induced focal derangements in liver metabolism.

biochemistry↗

bioRxiv2025

Decoding the causal drivers of spatial cellular topology

Decoding how cells influence and communicate with each other in space is fundamental for understanding tissue organization. However, existing approaches either overlook spatial context entirely or rely solely on local cell-cell adjacency, failing to capture how global tissue topology shapes cellular communication. Here, we present GLACIER, which introduces spatial Granger causality to infer transcriptional and signaling relationships that emerge from tissue organization. By combining GASTONs global isodepth coordinate with Veloramas graph-based causal inference framework, we enable bidirectional inference of regulatory relationships along spatial axes, identifying transcription factortarget interactions and ligand-receptor pairs that operate across spatial domains. Applying GLACIER to single-cell spatial transcriptomics data from the mouse cerebellum, we identify both continuous within-cell-type regulatory gradients and discontinuous drivers at layer interfaces, while distinguishing between forward and backward cellular communication along the isodepth axis. Our approach reveals how tissue architecture directs patterns of cellular communication, providing a framework for understanding spatially-encoded regulatory programs. Software availabilityhttps://github.com/rohitsinghlab/glacier.

bioinformatics↗

bioRxiv2025

GASTON-Mix: a unified model of spatial gradients and domains using spatial mixture-of-experts

MotivationGene expression varies across a tissue due to both the organization of the tissue into spatial domains, i.e. discrete regions of a tissue with distinct cell type composition, and continuous spatial gradients of gene expression within different spatial domains. Spatially resolved transcriptomics (SRT) technologies provide high-throughput measurements of gene expression in a tissue slice, enabling the characterization of spatial gradients and domains. However, existing computational methods for quantifying spatial variation in gene expression either model only spatial domains - and do not account for continuous gradients of expression - or require restrictive geometric assumptions on the spatial domains and spatial gradients that do not hold for many complex tissues. ResultsWe introduce GASTON-Mix, a machine learning algorithm to identify both spatial domains and spatial gradients within each domain from SRT data. GASTON-Mix extends the mixture-of-experts (MoE) deep learning framework to a spatial MoE model, combining the clustering component of the MoE model with a neural field model that learns a separate 1-D coordinate ("isodepth") within each domain. The spatial MoE is capable of representing any geometric arrangement of spatial domains in a tissue, and the isodepth coordinates define continuous gradients of gene expression within each domain. We show using simulations and real data that GASTON-Mix identifies spatial domains and spatial gradients of gene expression more accurately than existing methods. GASTON-Mix reveals spatial gradients in the striatum and lateral septum that regulate complex social behavior, and GASTON-Mix identifies localized spatial gradients of hypoxia and TNF- signaling in the tumor microenvironment.

bioinformatics↗

bioRxiv2024

Quantifying higher-order epistasis: beware the chimera

Epistasis, or interactions in which alleles at one locus modify the fitness effects of alleles at other loci, plays a fundamental role in genetics, protein evolution, and many other areas of biology. Epistasis is typically quantified by computing the deviation from the expected fitness under an additive or multiplicative model using one of several formulae. However, these formulae are not all equivalent. Importantly, one widely used formula - which we call the chimeric formula - measures deviations from a multiplicative fitness model on an additive scale, thus mixing two measurement scales. We show that for pairwise interactions, the chimeric formula yields a different magnitude, but the same sign (synergistic vs. antagonistic) of epistasis compared to the multiplicative formula that measures both fitness and deviations on a multiplicative scale. However, for higher-order interactions, we show that the chimeric formula can have both different magnitude and sign compared to the multiplicative formula -- thus confusing negative epistatic interactions with positive interactions, and vice versa. We resolve these inconsistencies by deriving fundamental connections between the different epistasis formulae and the parameters of the multivariate Bernoulli distribution. Our results demonstrate that the additive and multiplicative epistasis formulae are more mathematically sound than the chimeric formula. Moreover, we demonstrate that the mathematical issues with the chimeric epistasis formula lead to markedly different biological interpretations of real data. Analyzing multi-gene knockout data in yeast, multi-way drug interactions in E. coli, and deep mutational scanning (DMS) of several proteins, we find that 10 - 60% of higher-order interactions have a change in sign with the multiplicative or additive epistasis formula. These sign changes result in qualitatively different findings on functional divergence in the yeast genome, synergistic vs. antagonistic drug interactions, and and epistasis between protein mutations. In particular, in the yeast data, the more appropriate multiplicative formula identifies nearly 500 additional negative three-way interactions, thus extending the trigenic interaction network by 25%.

genetics↗

bioRxiv2024

A latent variable model for evaluating mutual exclusivity and co-occurrence between driver mutations in cancer

A key challenge in cancer genomics is understanding the functional relationships and dependencies between combinations of somatic mutations that drive cancer development. Such driver mutations frequently exhibit patterns of mutual exclusivity or co-occurrence across tumors, and many methods have been developed to identify such dependency patterns from bulk DNA sequencing data of a cohort of patients. However, while mutual exclusivity and co-occurrence are described as properties of driver mutations, existing methods do not explicitly disentangle functional, driver mutations from neutral, passenger mutations. In particular, nearly all existing methods evaluate mutual exclusivity or co-occurrence at the gene level, marking a gene as mutated if any mutation - driver or passenger - is present. Since some genes have a large number of passenger mutations, existing methods either restrict their analyses to a small subset of suspected driver genes - limiting their ability to identify novel dependencies - or make spurious inferences of mutual exclusivity and co-occurrence involving genes with many passenger mutations. We introduce DIALECT, an algorithm to identify dependencies between pairs of driver mutations from somatic mutation counts. We derive a latent variable mixture model for drivers and passengers that combines existing probabilistic models of passenger mutation rates with a latent variable describing the unknown status of a mutation as a driver or passenger. We use an expectation maximization (EM) algorithm to estimate the parameters of our model, including the rates of mutually exclusivity and co-occurrence between drivers. We demonstrate that DIALECT more accurately infers mutual exclusivity and co-occurrence between driver mutations compared to existing methods on both simulated mutation data and somatic mutation data from 5 cancer types in The Cancer Genome Atlas (TCGA).

cancer biology↗

bioRxiv2023

Mapping the topography of spatial gene expression with interpretable deep learning

Spatially resolved transcriptomics technologies provide high-throughput measurements of gene expression in a tissue slice, but the sparsity of this data complicates the analysis of spatial gene expression patterns such as gene expression gradients. We address these issues by deriving a topographic map of a tissue slice--analogous to a map of elevation in a landscape--using a novel quantity called the isodepth. Contours of constant isodepth enclose spatial domains with distinct cell type composition, while gradients of the isodepth indicate spatial directions of maximum change in gene expression. We develop GASTON, an unsupervised and interpretable deep learning algorithm that simultaneously learns the isodepth, spatial gene expression gradients, and piecewise linear functions of the isodepth that model both continuous gradients and discontinuous spatial variation in the expression of individual genes. We validate GASTON by showing that it accurately identifies spatial domains and marker genes across several biological systems. In SRT data from the brain, GASTON reveals gradients of neuronal differentiation and firing, and in SRT data from a tumor sample, GASTON infers gradients of metabolic activity and epithelial-mesenchymal transition (EMT)-related gene expression in the tumor microenvironment.

bioinformatics↗

bioRxiv2022

Belayer: Modeling discrete and continuous spatial variation in gene expression from spatially resolved transcriptomics

Spatially resolved transcriptomics (SRT) technologies measure gene expression at known locations in a tissue slice, enabling the identification of spatially varying genes or cell types. Current approaches for these tasks assume either that gene expression varies continuously across a tissue or that a slice contains a small number of regions with distinct cellular composition. We propose a model for SRT data that includes both continuous and discrete spatial variation in expression, and an algorithm, Belayer, to estimate the parameters of this model from layered tissues. Belayer models gene expression as a piecewise linear function of the relative depth of a tissue layer with possible discontinuities at layer boundaries. We use conformal maps to model relative depth and derive a dynamic programming algorithm to infer layer boundaries and gene expression functions. Belayer accurately identifies tissue layers and infers biologically meaningful spatially varying genes in SRT data from brain and skin tissue samples.

bioinformatics↗

bioRxiv2022

NetMix2: Unifying network propagation andaltered subnetworks

A standard paradigm in computational biology is to use interaction networks to analyze high-throughput biological data. Two common approaches for leveraging interaction networks are: (1) network ranking, where one ranks vertices in the network according to both vertex scores and network topology; (2) altered subnetwork identification, where one identifies one or more subnetworks in an interaction network using both vertex scores and network topology. The dominant approach in network ranking is network propagation which smooths vertex scores over the network using a random walk or diffusion process, thus utilizing the global structure of the network. For altered subnetwork identification, existing algorithms either restrict solutions to subnetworks in subnetwork families with simple topological constraints, such as connected subnetworks, or utilize ad hoc heuristics that lack a rigorous statistical foundation. In this work, we unify the network propagation and altered subnetwork approaches. We derive a subnetwork family which we call the propagation family that approximates the subnetworks ranked highly by network propagation. We introduce NetMix2, a principled algorithm for identifying altered subnetworks from a wide range of subnetwork families, including the propagation family, thus combining the advantages of the network propagation and altered subnetwork approaches. We show that NetMix2 outperforms network propagation on data simulated using the propagation family. Furthermore, NetMix2 outperforms other methods at recovering known disease genes in pan-cancer somatic mutation data and in genome-wide association data from multiple human diseases. NetMix2 is publicly available at https://github.com/raphael-group/netmix2.

bioinformatics↗

bioRxiv2020

NetMix: A network-structured mixture model for reduced-bias estimation of altered subnetworks

A classic problem in computational biology is the identification of altered subnetworks: subnetworks of an interaction network that contain genes/proteins that are differentially expressed, highly mutated, or otherwise aberrant compared to other genes/proteins. Numerous methods have been developed to solve this problem under various assumptions, but the statistical properties of these methods are often unknown. For example, some widely-used methods are reported to output very large subnetworks that are difficult to interpret biologically. In this work, we formulate the identification of altered subnetworks as the problem of estimating the parameters of a class of probability distributions which we call the Altered Subset Distribution (ASD). We derive a connection between a popular method, jActiveModules, and the maximum likelihood estimator (MLE) of the ASD. We show that the MLE is statistically biased, explaining the large subnetworks output by jActiveModules. We introduce NetMix, an algorithm that uses Gaussian mixture models to obtain less biased estimates of the parameters of the ASD. We demonstrate that NetMix outperforms existing methods in identifying altered subnetworks on both simulated and real data, including the identification of differentially expressed genes from both microarray and RNA-seq experiments and the identification of cancer driver genes in somatic mutation data. AvailabilityNetMix is available online at https://github.com/raphael-group/netmix. Contactbraphael@princeton.edu

bioinformatics↗