ABSTRACT:The regulation of the switch from fetal (HBG) to adult (HBB and HBD) β-globin gene expression has served as a paradigm for clinically relevant developmental transcriptional control. Mechanistic studies of this switch have predominantly focused on HBG repressors, with comparatively little attention paid to potential HBG activators. We found that in adult-type HUDEP2 erythroid cells, the ATP-dependent chromatin remodeler Brahma Related Gene 1 (BRG1) preferentially activates the HBG genes as well as the minor adult HBD gene. BRG1 is a core catalytic subunit of 3 BRG1/BRM-associated factor (BAF) complexes, canonical BAF, polybromo BAF, and noncanonical BAF (ncBAF) that regulate chromatin accessibility in distinct gene- and cell-type contexts. To dissect the specific BAF complex configuration mediating selective activation of HBG and HBD in erythroid cells, we performed CRISPR-mediated targeting of individual subunits and pinpointed the regulatory activity to the ncBAF complex. Loss of the ncBAF complex subunits BRD9 and BAF60A preferentially decreased HBG and HBD transcription while accelerating terminal erythroid differentiation and hemoglobinization. Acute pharmacological depletion of BRD9 in HUDEP2 and primary erythroid cells selectively reduced transcription of HBD and HBG, suggesting direct effects at these genes. Collectively, our unexpected findings demonstrate that the BAF complex, through distinct subcomplex configurations, can regulate selective gene expression within a multigene cluster. This expands the traditional view of BAF as a general coactivator, highlights its role in gene-specific regulation, and identifies a potential target for therapeutic manipulation of β-like globin genes in erythroid cell disorders.

Reactivating the fetal globin genes HBG1 and HBG2 in adult erythroid cells represents a validated therapeutic approach for hemoglobinopathies. Central mediators of the fetal-to-adult hemoglobin transition include the direct transcriptional HBG1/2 repressors BCL11A1,2, LRF3, and NFIA/X4. Limited-scale screens have attempted to expand the regulatory circuity surrounding fetal globin silencing, but systematic genome-wide dissection of such pathways is lacking. We employed a two-tiered genetic screening strategy - a novel CRISPR-Cas12a-based screening platform followed by a domain-focused CRISPR-Cas9 screen - to interrogate all known human coding genes for their impact on HBG1/2 regulation and erythroid cellular fitness, generating a comprehensive resource for the field. Among the top hits was PTPA, an activator of the serine-threonine phosphatase PP2A whose loss elevates HBG1/2 levels while preserving erythroid differentiation. Phenotypic rescue experiments revealed that PTPA silences HBG1/2 expression primarily by regulating BCL11A expression. To our knowledge, this study represents the most comprehensive CRISPR dissection of HBG regulation to date, highlighting the power of Cas12a-based genome-scale screening for uncovering disease-relevant pathways.

Spatial enhancer connectivity is fundamental to proper gene regulation. Enhancer dysregulation has emerged as a hallmark of cancers, including T-cell acute lymphoblastic leukemias (T-ALL). T-ALL are aggressive malignancies characterized by marked transcriptional heterogeneity driven by distinct stages of developmental arrest and diverse noncoding alterations. How these cancers co-opt nuclear architecture to rewire enhancer connectivity remains poorly understood. Here, we report that the LDB1 chromatin architectural complex is an essential mediator of enhancer-oncogene looping that sustains oncogenic transcriptional programs across multiple T-ALL subtypes. Integrating bulk and single-cell transcriptomic data from patients with T-ALL and healthy hematopoietic controls, we show that the LDB1-dependent regulatory circuitry defines the molecular identities of distinct T-ALL subtypes while restricting plasticity toward alternative cell states. LDB1 loss dismantles chromatin looping among cell state-defining enhancers liberating them to form promiscuous interactions with nearby genes. This enhancer rewiring stimulates expression of key metabolic genes, creating a mevalonate pathway dependency exploitable with statin treatment. Our study establishes LDB1 as a central executor of T-ALL regulatory circuitry and more broadly illustrates chromatin rewiring as a source of targetable dependencies in cancer.

ABSTRACT:Polycomb repressive complex 1 (PRC1) and PRC2 regulate diverse developmental processes, including the fetal-to-adult switch in hemoglobin (Hb) production, a process whose reversal is a goal for the treatment of sickle cell disease and β-thalassemia. PRC inhibitors show promise for various disorders, but use is limited because of pleiotropic PRC activities. We explored whether fetal Hb (HbF) can be reactivated in adult erythroid cells by selective perturbations of PRC1 or PRC2 components without complete loss of PRC function. A high-density CRISPR-CRISPR-associated protein 9 (Cas9) mutagenesis screen identified a region in EZH2 in which Cas9 induced exon 14 skipping (EZH2Δ14). EZH2Δ14, which lacks a portion of the CXC domain, relieves HbF repression while largely maintaining cellular fitness. EZH2Δ14 retains H3K27 methylation and repression of a PRC target gene subset. Experiments in cells derived from mice bearing human β-globin genes confirm that pathways mediating EZH2 control of HbF expression can function in a mouse model of HBG switching. These findings demonstrate that partial disruption of PRC can yield selective phenotypes, highlighting the therapeutic potential of targeting nonenzymatic domains within chromatin-modifying complexes.

Abstract LDB1 and LMO2, two proteins frequently overexpressed in T-cell acute lymphoblastic leukemia (T-ALL), form a chromatin architectural complex that promotes chromatin looping between enhancers and/or promoters. Here, we defined LDB1-driven oncogenic enhancer connectivity in T-ALL and examined its impact on therapeutic vulnerabilities. To identify proximal LDB1 targets in T-ALL, we engineered isogenic T-ALL cell lines (LOUCY [ETP-ALL] and KOPT K1 [non-ETP-ALL]) with dTAG degrons at the endogenous LDB1 loci. Treatment with dTAG-V1 ligand for 4 hours reduced LDB1 protein levels and chromatin occupancy by >90%. Acute LDB1 loss disrupted spatial enhancer connectivity at critical leukemic oncogenes in both LOUCY (e.g. HHEX, MYB, MYCN) and KOPT K1 (e.g. DUSP6, STAT4), resulting in their downregulation and subsequent reduction in cell growth. CRISPRa-mediated restoration of select LDB1-dependent oncogenes, such as MYB, in LDB1-depleted cells rescued cell expansion. Nascent transcript profiling further revealed that acute LDB1 loss affected distinct gene sets in LOUCY and KOPT K1, consistent with their distinct identities. Notably, however, several cholesterol biosynthetic genes (HMGCS1, MVD, MVK) were upregulated in both cell types. These changes were not driven by altered expression or subcellular localization of SREBP2—the canonical transcriptional regulator of these genes. Instead, they were caused by altered enhancer-promoter connectivity in the absence of LDB1. Specifically, LDB1 loss disrupted connectivity of nearby enhancers for different genes, thereby liberating them to form de novo contacts with HMGCS1, MVK, and MVD gene promoters to activate them. Hence, by clustering regulatory elements, LDB1 not only enables the expression of genes but also constrains enhancers from making inappropriate contacts. Importantly, the native LDB1-dependent loops are detected in primary samples from patients with T-ALL based on published H3K27ac HiChIP data. Functionally, we found that cholesterol flux and expression of upstream regulators, like SREBP2, increased with prolonged LDB1 depletion, suggesting a feed-forward regulatory mechanism on cholesterol homeostasis in T-ALL. Using both LDB1 degron and knockout models, we observed that LDB1 loss sensitizes leukemic cells to rosuvastatin and pitavastatin by 2-5-fold. This was recapitulated in several T-ALL cell lines with distinct molecular drivers upon LDB1 knockout. Furthermore, statin sensitization was rescued by spike-in of mevalonate or geranylgeranyl pyrophosphate, confirming that this phenotype is driven by altered cholesterol metabolism. Together, our study illustrates a paradigm by which LDB1 loss enables illegitimate spatial connections of enhancers with cholesterol biosynthetic gene promoters. This, in turn, creates a new metabolic addiction in leukemic cells, which may be targeted with statins. Citation Format: Rahul S. Bhansali, Juan S. Long, Siqing Wang, Ahnaf Tausif, Shuo Zhang, Petri Pölönen, Sarah Skuli, Nicholas Aboreden, Zhuangzhuang Geng, Belinda M. Giardine, Cheryl A. Keller, Ross C. Hardison, Charles G. Mullighan, Gerd A. Blobel. LDB1-dependent enhancer connectivity constrains a metabolic synthetic lethality in T-cell acute lymphoblastic leukemia [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 1 (Regular Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(7 Suppl):Abstract nr 7237.

Studies of globin gene clusters have established many paradigms of gene regulation. This review focuses on the α- and β-globin gene clusters of humans and mice, summarizing important insights from high-throughput biochemical assays and directed genetic dissections and emphasizing similarities across the types of gene clusters and between species. The overall arrangements and architectures are similar, with each gene cluster being localized within a topologically constrained unit of chromatin containing a multicomponent enhancer (i.e., a locus control region) and other regulatory elements bound by a similar set of transcription factors and coactivators. Differential expression of the globin genes within each cluster during ontogeny is associated with changes in contacts with the locus control region and involves the action of gene-specific repressors. Detailed study of the fetal β-like HBG1 and HBG2 globin genes has revealed a remarkable diversity of regulatory pathways that provide candidates for therapeutic approaches to reactivate these genes for β-hemoglobinopathies.

How specific enhancer-promoter pairing is established remains mostly unclear. Besides the CTCF/cohesin machinery, few nuclear factors have been studied for a direct role in physically connecting regulatory elements. Using a murine erythroid cell model, we show via acute degradation experiments that LDB1 directly and broadly promotes connectivity among regulatory elements. Most LDB1-mediated contacts, even those spanning hundreds of kb, can form in the absence of CTCF, cohesin, or YY1 as determined using multiple degron systems. Moreover, an engineered LDB1-driven chromatin loop is cohesin independent. Cohesin-driven loop extrusion does not stall at LDB1-occupied sites but aids the formation of a subset of LDB1-anchored loops. Leveraging the dynamic reorganization of nuclear architecture during the transition from mitosis to G1 phase, we observe that loop formation and de novo LDB1 occupancy correlate and can occur independently of structural loops. Tri-C and Region Capture Micro-C reveal that LDB1 organizes multi-enhancer networks to activate transcription. These findings establish LDB1 as a driver of spatial connectivity.

Changes in gene expression drive differentiation along distinct cell lineages, and these shifts in gene expression are associated with alterations in chromatin accessibility and modifications reflecting activation or repression. We used deep sequencing of polyA+ RNA to map the transcriptomes of the megakaryocyte-erythroid progenitor (MEP) and cells of its two daughter lineages, erythroblasts (ERY) and megakaryocytes (MEG) in mice to reveal insights into differentiation. Transcriptome comparisons revealed that MEPs already expressed much of the MEG program while continuing to express genes associated with parallel myeloid lineages. By contrast, ERY underwent an extensive program of gene induction along with repression of pan-hematopoietic and MEG genes. Maps of transcription factor (TF) occupancy also indicated distinct modes of regulation for the MEG and ERY programs, with MEG genes preferentially occupied by hematopoietic TFs in multipotent progenitors and continued occupancy post-commitment, in contrast to erythroid genes that were primarily occupied in committed ERY. Previous work had indicated a surprising discordance in the clustering of MEP with other hematopoietic cell types by RNA-seq versus chromatin states. We combined the differential expression data with chromatin accessibility across blood cell types to identify trends that contribute to this discordance. Specifically, candidate cis-regulatory elements (cCREs) in some ERY-specific genes were precociously actuated in the bipotential cell populations, and some other genes were expressed in both the MEP population and MEG but their cCREs have less chromatin accessibility in MEP. This discordance in cell type clustering by different modalities of functional genomics may reflect the different contributions of subpopulations in the MEP to the different modalities measured.

CpG islands in vertebrate genomes are associated not only with housekeeping genes but also with some tissue-restricted genes, including those that comprise the alpha-like globin gene clusters of humans and rabbits. The facts that CpG islands are unmethylated in all tissues (even those not expressing the genes) and that they adopt a unique chromatin structure may have influenced the evolution of mechanisms by which alpha-globin genes are regulated to achieve expression in the correct tissues and the proper developmental stage. We have shown that CpG islands from the alpha-globin gene cluster have positive effects on gene expression, which requires integration into chromosomes and hence is likely to be mediated by effects on chromatin. The potential for a DNA sequence to form a nucleosome complex with core histones is influenced by the topology of that DNA molecule in solution. In light of the unique sequence profile of CpG islands, we tested whether CpG island DNA has a different affinity for nucleosome formation than non-CpG island DNA in a competitive nucleosome reconstitution assay. We find that CpG island DNA from the rabbit alpha-globin gene has a relatively low affinity for core histones in this system. This lower affinity was observed regardless of the position of the fragment in the gene or 5' flanking region. The lower affinity for CpG island DNA showed a length dependence, with the relative affinity for histones decreasing for longer DNA fragments of CpG island DNA. The results are consistent with predictions based on the functional effects of CpG islands and emerging rules for sequences that can exclude nucleosomes.

Transcriptional enhancers are brought into proximity with promoters via chromatin looping. The architectural transcription cofactor LDB1 facilitates spatial connectivity among enhancers and promoters, but whether this occurs through simple dimerization or requires heterotypic protein assemblies is unknown. Here, we investigated the role of single-stranded DNA-binding proteins (SSBPs) in regulating LDB1-mediated chromatin looping and transcription. SSBP2, SSBP3, and SSBP4 colocalize with LDB1 genome wide. Among these, only SSBP3 is essential for murine erythroid cell viability, LDB1 function, and transcription. LDB1, but not single-stranded DNA, is the predominant genome-wide tether of SSBP3 to chromatin. SSBP3 depletion in SSBP2/4 knockout cells globally weakened LDB1-dependent chromatin loops and lowered nascent transcription without impacting LDB1's chromatin binding. Chromatin tethering experiments revealed that SSBP3 and LDB1 mutually depend on each other to form looped contacts. SSBP3 stabilizes LDB1 dimerization in vitro, providing a possible mechanism. SSBPs emerge as key functional components of the architectural LDB1 complex.

Topologically associating domains (TADs) and stripes are important architectural structures on Hi-C data that are important for gene regulation. We present Joint Optimized nested TADs and Stripes (JOnTADS), a unified caller for TADs and stripes in Hi-C data. JOnTADS effectively identifies hierarchical TADs and stripes in population Hi-C and micro-C datasets, and hierarchical TADs in single-cell Hi-C data. It provides robust identifications aligned with known biology and effectively captures interaction frequency variations in contact maps across diverse contexts. When multiple samples are available, JOnTADS leverages shared information across samples in TAD identification, reducing unwanted variation in TAD boundary identification while preserving biological differences. This approach enables robust identifications in single-cell Hi-C data, effectively addressing challenges posed by data sparsity. JOnTADS is computationally efficient and requires minimal user tuning. ### Competing Interest Statement The authors have declared no competing interest.

Despite the discovery of LMO2 translocations in T-cell acute lymphoblastic leukemia (T-ALL) over 30 years ago, the mechanism by which LMO2 contributes to leukemogenesis remains unclear. LMO2 acts as a molecular adaptor bringing the chromatin architectural protein LDB1 into proximity with DNA-binding proteins such as GATA and basic helix-loop-helix transcription factors like TAL1. Our group and others have previously demonstrated that LDB1 is both necessary and sufficient to form long range chromatin loops between enhancers and promoters. Previous studies have found that overexpression of Lmo2 in transgenic mouse thymocytes leads to the development of T-ALL with high penetrance in a manner dependent on LDB1. Thus, we hypothesized that LDB1 aberrantly forms enhancer-promoter loops at oncogenic loci in LMO2-overexpressing forms of T-ALL, thereby driving an oncogenic transcriptional program. To elucidate the transcriptional network regulated by LDB1, we first profiled LDB1 occupancy in two T-ALL cell lines with high LMO2 expression: LOUCY, an early thymic precursor ALL, and KOPT K1, a near late cortical T-ALL. ChIP-seq demonstrated that the majority of LDB1 peaks fall within enhancer regions. Surprisingly, over 60% LDB1-occupied regions in LOUCY do not overlap with those observed in KOPT K1. Motif enrichment analysis at non-overlapping enhancers in LOUCY also displayed preferential enrichment for HOX and RUNX motifs while those in KOPT K1 were enriched for STAT and NFAT, suggesting that transcriptional miswiring in different T-ALL subtypes may rely on distinct enhancer compositions. To study the direct effects of LDB1 on gene expression and chromatin folding, we generated isogenic LOUCY and KOPT K1 cell lines harboring homozygous knock-ins of a FKBP12F36V degron (dTAG), enabling us to degrade LDB1 almost entirely upon dTAG-V1 ligand exposure for 4 hours. Importantly, we did not observe significant changes in chromatin accessibility (ATAC-seq) or enhancer activity (H3K27ac ChIP-seq) upon acute LDB1 degradation, suggesting that architectural and transcriptional changes are attributable to LDB1 loss as opposed to secondary changes due to lost enhancer activity. Furthermore, LMO2 co-occupancy with LDB1 remained relatively unperturbed at these sites further underscoring LDB1 as a direct mediator of chromatin looping. We performed Micro-C to evaluate chromatin architecture acutely dependent on LDB1. Chromatin compartments and topologically associated domains (TADs) remained intact upon LDB1 degradation. However, we observed widespread changes in chromatin loops, particularly those involving enhancer-promoter and enhancer-enhancer contacts that will be discussed in detail. We next used TT-seq to assess changes in nascent transcription upon LDB1 depletion. Genes regulated by LDB1 in both T-ALL subtypes are the BCL2 family members BCL2L1 and MCL1. The BCL2 family members represent critical anti-apoptotic proteins which are often upregulated in hematologic malignancies and are part of a therapeutic axis sensitive to BH3 mimetic agents. We confirmed that acute degradation of LDB1 leads to a decrease in nascent transcription of these genes, indicating that their expression is in part dependent on LDB1. Despite the commonality of this pathway in both T-ALL subtypes, we also observed distinct patterns of LDB1-mediated transcriptional regulation unique to each cell line associated with developmentally restricted genes. Notably, persistent LDB1 depletion leads to restricted cell growth of LOUCY cells but not KOPT K1 cells, underscoring the differences in the transcriptional programs regulated in these two cell types. Together, these data suggest that LDB1 is a critical transcriptional regulator in LMO2-high T-ALL through regulation of enhancer-promoter looping. Acute perturbation of LDB1-dependent enhancer-promoter connectivity provides novel insights into T-ALL biology and the regulation of genes associated with potentially targetable pathways.

Few transcription factors have been examined for their direct roles in physically connecting enhancers and promoters. Here acute degradation of Yin Yang 1 (YY1) in erythroid cells revealed its requirement for the maintenance of numerous enhancer-promoter loops, but not compartments or domains. Despite its reported ability to interact with cohesin, the formation of YY1-dependent enhancer-promoter loops does not involve stalling of cohesin-mediated loop extrusion. Integrating mitosis-to-G1-phase dynamics, we observed partial retention of YY1 on mitotic chromatin, predominantly at gene promoters, followed by rapid rebinding during mitotic exit, coinciding with enhancer-promoter loop establishment. YY1 degradation during the mitosis-to-G1-phase interval revealed a set of enhancer-promoter loops that require YY1 for establishment during G1-phase entry but not for maintenance in interphase, suggesting that cell cycle stage influences YY1's architectural function. Thus, as revealed here for YY1, chromatin architectural functions of transcription factors can vary in their interplay with CTCF and cohesin as well as by cell cycle stage. Yin Yang 1 (YY1) aids in the formation of enhancer-promoter (E-P) loops independently of cohesin. YY1 maintains a subset of E-P interactions in interphase and establishes an overlapping yet distinct set after mitotic exit.

The switch from fetal hemoglobin ( gamma- globin, HBG ) to adult hemoglobin ( beta- globin, HBB gene transcription in erythroid cells serves as a paradigm for a complex and clinically relevant developmental gene regulatory program. We previously identi fi ed HIC2 as regulator of the switch by inhibiting the transcription of BCL11A , a key repressor of HBG production. HIC2 is highly expressed in fetal cells, but the mechanism of its regulation unclear. Here we report that HIC2 developmental expression is controlled by microRNAs (miRNAs), as loss of global miRNA biogenesis through DICER1 depletion leads to upregulation of HIC2 and HBG messenger RNA. We identi fi ed the adult -expressed let miRNA family as a direct posttranscriptional regulator of HIC2 . Ectopic expression of let in fetal cells lowered HIC2 levels, whereas inhibition of let -7 in adult erythroblasts increased HIC2 production, culminating in decommissioning of a BCL11A erythroid enhancer and reduced BCL11A transcription. HIC2 depletion let -7 -inhibited cells restored BCL11A-mediated repression of HBG . Together, these data establish that fetal hemoglobin silencing in adult erythroid cells is under the control of a miRNA-mediated inhibitory pathway ( let -7 -I HIC2 BCL11A -I HBG ).

ABSTRACTDuring mitosis, condensin activity interferes with interphase chromatin structures. Here, we generated condensin-free mitotic chromosomes to investigate genome folding principles. Co- depletion of condensin I and II, but neither alone, triggered mitotic chromosome compartmentalization in ways that differ from interphase. Two distinct euchromatic compartments, indistinguishable in interphase, rapidly emerged upon condensin loss with different interaction preferences and dependence on H3K27ac. Constitutive heterochromatin gradually self-aggregated and co-compartmentalized with the facultative heterochromatin, contrasting with their separation during interphase. While topologically associating domains (TADs) and CTCF/cohesin mediated structural loops remained undetectable, cis-regulatory element contacts became apparent, providing an explanation for their quick re-establishment during mitotic exit. HP1 proteins, which are thought to partition constitutive heterochromatin, were absent from mitotic chromosomes, suggesting, surprisingly, that constitutive heterochromatin can self-aggregate without HP1. Indeed, in cells traversing from M- to G1-phase in the combined absence of HP1α, HP1Π and HP1γ, re-established constitutive heterochromatin compartments normally. In sum, “clean-slate” condensin-deficient mitotic chromosomes illuminate mechanisms of genome compartmentalization not revealed in interphase cells.

Knowledge of locations and activities of cis-regulatory elements (CREs) is needed to decipher basic mechanisms of gene regulation and to understand the impact of genetic variants on complex traits. Previous studies identified candidate CREs (cCREs) using epigenetic features in one species, making comparisons difficult between species. In contrast, we conducted an interspecies study defining epigenetic states and identifying cCREs in blood cell types to generate regulatory maps that are comparable between species, using integrative modeling of eight epigenetic features jointly in human and mouse in our Validated Systematic Integration (VISION) Project. The resulting catalogs of cCREs are useful resources for further studies of gene regulation in blood cells, indicated by high overlap with known functional elements and strong enrichment for human genetic variants associated with blood cell phenotypes. The contribution of each epigenetic state in cCREs to gene regulation, inferred from a multivariate regression, was used to estimate epigenetic state regulatory potential (esRP) scores for each cCRE in each cell type, which were used to categorize dynamic changes in cCREs. Groups of cCREs displaying similar patterns of regulatory activity in human and mouse cell types, obtained by joint clustering on esRP scores, harbor distinctive transcription factor binding motifs that are similar between species. An interspecies comparison of cCREs revealed both conserved and species-specific patterns of epigenetic evolution. Finally, we show that comparisons of the epigenetic landscape between species can reveal elements with similar roles in regulation, even in the absence of genomic sequence alignment.

Transcriptional enhancers have been extensively characterized, but cis-regulatory elements involved in acute gene repression have received less attention. Transcription factor GATA1 promotes erythroid differentiation by activating and repressing distinct gene sets. Here, we study the mechanism by which GATA1 silences the proliferative gene Kit during murine erythroid cell maturation and define stages from initial loss of activation to heterochromatinization. We find that GATA1 inactivates a potent upstream enhancer but concomitantly creates a discrete intronic regulatory region marked by H3K27ac, short noncoding RNAs, and de novo chromatin looping. This enhancer-like element forms transiently and serves to delay Kit silencing. The element is ultimately erased via the FOG1/NuRD deacetylase complex, as revealed by the study of a disease-associated GATA1 variant. Hence, regulatory sites can be self-limiting by dynamic co-factor usage. Genome-wide analyses across cell types and species uncover transiently active elements at numerous genes during repression, suggesting that modulation of silencing kinetics is widespread.

Spatial transcriptomics (ST) profiles gene expression in intact tissues. However, ST data measured at each spatial location may represent gene expression of multiple cell types, making it difficult to identify cell-type-specific transcriptional variation across spatial contexts. Existing cell-type deconvolutions of ST data often require single-cell transcriptomic references, which can be limited by availability, completeness and platform effect of such references. We present RETROFIT, a reference-free Bayesian method that produces sparse and interpretable solutions to deconvolve cell types underlying each location independent of single-cell transcriptomic references. Results from synthetic and real ST datasets acquired by Slide-seq and Visium platforms demonstrate that RETROFIT outperforms existing reference-based and reference-free methods in estimating cell-type composition and reconstructing gene expression. Applying RETROFIT to human intestinal development ST data reveals spatiotemporal patterns of cellular composition and transcriptional specificity. RETROFIT is available at https://bioconductor.org/packages/release/bioc/html/retrofit.html.

BACKGROUND:Epigenetic modification of chromatin plays a pivotal role in regulating gene expression during cell differentiation. The scale and complexity of epigenetic data pose significant challenges for biologists to identify the regulatory events controlling cell differentiation. RESULTS:To reduce the complexity, we developed a package, called Snapshot, for clustering and visualizing candidate cis-regulatory elements (cCREs) based on their epigenetic signals during cell differentiation. This package first introduces a binarized indexing strategy for clustering the cCREs. It then provides a series of easily interpretable figures for visualizing the signal and epigenetic state patterns of the cCREs clusters during the cell differentiation. It can also use different hierarchies of cell types to highlight the epigenetic history specific to any particular cell lineage. We demonstrate the utility of Snapshot using data from a consortium project for ValIdated Systematic IntegratiON (VISION) of epigenomic data in hematopoiesis. CONCLUSION:The package Snapshot can identify all distinct clusters of genomic locations with unique epigenetic signal patterns during cell differentiation. It outperforms other methods in terms of interpreting and reproducing the identified cCREs clusters. The package of Snapshot is available at GitHub: https://github.com/guanjue/Snapshot .