Aaron Streets
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Aaron Streets has a total of 342 co-authors. Here are the top co-authors by number of shared publications:
- Nicolas Altemose
Department of Genetics, Stanford University, Palo Alto, USA
15 shared publications - Nir Yosef
Department of Systems Immunology, Weizmann Institute of Science, Rehovot, Israel
15 shared publications - Adam Gayoso
Google DeepMind
9 shared publications - Aaron F. Straight
Stanford University
8 shared publications - Annie Maslan
UC Berkeley–UCSF Graduate Program in Bioengineering, University of California, Berkeley, California 94720, USA
8 shared publications - Jonathan A. White
Division of Biology and Biological Engineering, California Institute of Technology, Pasadena, CA, USA
8 shared publications
Latest Publications
- Open Access
Denoising single-cell ATAC-seq data with Shannon entropy
Single-cell Assay for Transposase Accessible Chromatin with sequencing (scATAC-seq) is widely used to investigate chromatin accessibility but experiments often contain empty droplets, droplets containing damaged nuclei or cellular-debris. Accurately filtering high-quality nuclei is important as noisy droplets contribute reads that can distort downstream analysis and obscure rare cell population. Current approaches often rely on read depth and or predefined genomic features that are not exactly the same as chromatin accessibility landscape. Here, we present a Shannon entropy-based method that filters high-quality nuclei more robustly by quantifying genome-wide accessibility complexity. This approach improves signal-to-noise ratio, increases sensitivity for accessible-region detection, and enables more reliable downstream inference in scATAC-seq studies.
Open AccessDenoising single-cell ATAC-seq data with Shannon entropy
Single-cell Assay for Transposase Accessible Chromatin with sequencing (scATAC-seq) is widely used to investigate chromatin accessibility but experiments often contain empty droplets, droplets containing damaged nuclei or cellular-debris. Accurately filtering high-quality nuclei is important as noisy droplets contribute reads that can distort downstream analysis and obscure rare cell population. Current approaches often rely on read depth and or predefined genomic features that are not exactly the same as chromatin accessibility landscape. Here, we present a Shannon entropy-based method that filters high-quality nuclei more robustly by quantifying genome-wide accessibility complexity. This approach improves signal-to-noise ratio, increases sensitivity for accessible-region detection, and enables more reliable downstream inference in scATAC-seq studies.
- Open Access
- 5 August 2026
- Published by Springer Nature in Nature Communications
- Vol. 17 (1) , 9471
Single-nucleus analysis of human white adipose tissue reveals adipocyte subsets with distinct metabolic profiles
Anatomic location of white adipose tissue is a determinant of cardiometabolic risk. To understand differences within/between adipose depots, we generated 65,668 single-nucleus transcriptomes from human subcutaneous or intraabdominal adipose tissue (SAT/IAT). Unsupervised analysis reveals 26 adipose-resident cell clusters including two subpopulations of mature adipocytes, characterized by high vs. low expression of adipocyte maturation genes (ADIPOMAThi vs. ADIPOMATlo). ADIPOMATlo adipocytes demonstrate a low-differentiation, pro-inflammatory, and pro-fibrotic transcriptome. IAT-resident ADIPOMATlo were more abundant in higher BMI donors, while SAT-resident ADIPOMATlo associated with impaired glycemia. TSHZ3 is identified as a candidate regulator of ADIPOMATlo transcriptome. TSHZ3 knockdown in adipogenic progenitors inhibits differentiation, with downregulation of early adipogenic regulators (e.g., CEBPA/B, PPARG) and mature adipocyte genes. Heterozygous deletion of Tshz3 in mice reduces SAT and IAT weight. Here, we show that adipocyte subsets with distinct transcriptomic signature reside in human WAT; altered TSHZ3-mediated transcriptional regulation may contribute to low-maturation subpopulation linked to metabolic disease.
- 5 August 2026
- Published by Springer Nature in Nature Communications
- Vol. 17 (1) , 9471
Open AccessSingle-nucleus analysis of human white adipose tissue reveals adipocyte subsets with distinct metabolic profiles
Anatomic location of white adipose tissue is a determinant of cardiometabolic risk. To understand differences within/between adipose depots, we generated 65,668 single-nucleus transcriptomes from human subcutaneous or intraabdominal adipose tissue (SAT/IAT). Unsupervised analysis reveals 26 adipose-resident cell clusters including two subpopulations of mature adipocytes, characterized by high vs. low expression of adipocyte maturation genes (ADIPOMAThi vs. ADIPOMATlo). ADIPOMATlo adipocytes demonstrate a low-differentiation, pro-inflammatory, and pro-fibrotic transcriptome. IAT-resident ADIPOMATlo were more abundant in higher BMI donors, while SAT-resident ADIPOMATlo associated with impaired glycemia. TSHZ3 is identified as a candidate regulator of ADIPOMATlo transcriptome. TSHZ3 knockdown in adipogenic progenitors inhibits differentiation, with downregulation of early adipogenic regulators (e.g., CEBPA/B, PPARG) and mature adipocyte genes. Heterozygous deletion of Tshz3 in mice reduces SAT and IAT weight. Here, we show that adipocyte subsets with distinct transcriptomic signature reside in human WAT; altered TSHZ3-mediated transcriptional regulation may contribute to low-maturation subpopulation linked to metabolic disease.
- Open Access
Epigenetic conditioning improves sequence-based modeling of gene regulation across cell types and alleles
Epigenetic state modulates gene regulation in a manner not always predictable from DNA sequence alone, yet current genomic deep learning models do not leverage epigenetic state as input. We present MethylSeqNet, a model that conditions pretrained sequence embeddings on CpG methylation, a stable epigenetic mark increasingly available from long-read sequencing data. Using a novel conditioning mechanism enabling scalability and interpretability, MethylSeqNet improves predictions in cases where differential epigenetic state drives regulatory variation. We show improvements over a sequence-only baseline for cell-type-specific chromatin accessibility and transcription. Epigenetic conditioning enables prediction of phenomena not encoded in allele sequence, including parent-of-origin imprinting, random monoallelic activity, and X-inactivation. We highlight a promising application of methylation conditioning by predicting the effects of a structural rearrangement in one rare disease patient case study. In silico motif insertion analysis confirms that MethylSeqNet learns methylation-dependent regulatory grammar, establishing a paradigm for integrating epigenetic information into genomic deep learning with immediate applications in rare disease interpretation.
Open AccessEpigenetic conditioning improves sequence-based modeling of gene regulation across cell types and alleles
Epigenetic state modulates gene regulation in a manner not always predictable from DNA sequence alone, yet current genomic deep learning models do not leverage epigenetic state as input. We present MethylSeqNet, a model that conditions pretrained sequence embeddings on CpG methylation, a stable epigenetic mark increasingly available from long-read sequencing data. Using a novel conditioning mechanism enabling scalability and interpretability, MethylSeqNet improves predictions in cases where differential epigenetic state drives regulatory variation. We show improvements over a sequence-only baseline for cell-type-specific chromatin accessibility and transcription. Epigenetic conditioning enables prediction of phenomena not encoded in allele sequence, including parent-of-origin imprinting, random monoallelic activity, and X-inactivation. We highlight a promising application of methylation conditioning by predicting the effects of a structural rearrangement in one rare disease patient case study. In silico motif insertion analysis confirms that MethylSeqNet learns methylation-dependent regulatory grammar, establishing a paradigm for integrating epigenetic information into genomic deep learning with immediate applications in rare disease interpretation.
- Open Access
Verifying LLM-extracted text with token alignment
Large language models excel at text extraction, but they sometimes hallucinate. A simple way to avoid hallucinations is to remove any extracted text that does not appear in the original source. This is easy when the extracted text is contiguous (findable with exact string matching), but much harder when it is discontiguous. Techniques for finding discontiguous phrases depend heavily on how the text is split—i.e., how it is tokenized. In this study, we show that splitting text along subword boundaries, with LLM-specific tokenizers, and aligning extracted text with ordered alignment algorithms, improves alignment by about 50% compared to word-level tokenization. To demonstrate this, we introduce the Berkeley Ordered Alignment of Text ( BOAT ) dataset, a modification of the Stanford Question Answering Dataset ( SQuAD ) that includes non-contiguous phrases, and BIO-BOAT a biomedical variant built from 51 bioRxiv preprints. We show that text-alignment methods form a partially ordered set, and that ordered alignment is the most practical choice for verifying LLM-extracted text. We implement this approach in taln , which enumerates ordinal subword alignments.
Open AccessVerifying LLM-extracted text with token alignment
Large language models excel at text extraction, but they sometimes hallucinate. A simple way to avoid hallucinations is to remove any extracted text that does not appear in the original source. This is easy when the extracted text is contiguous (findable with exact string matching), but much harder when it is discontiguous. Techniques for finding discontiguous phrases depend heavily on how the text is split—i.e., how it is tokenized. In this study, we show that splitting text along subword boundaries, with LLM-specific tokenizers, and aligning extracted text with ordered alignment algorithms, improves alignment by about 50% compared to word-level tokenization. To demonstrate this, we introduce the Berkeley Ordered Alignment of Text ( BOAT ) dataset, a modification of the Stanford Question Answering Dataset ( SQuAD ) that includes non-contiguous phrases, and BIO-BOAT a biomedical variant built from 51 bioRxiv preprints. We show that text-alignment methods form a partially ordered set, and that ordered alignment is the most practical choice for verifying LLM-extracted text. We implement this approach in taln , which enumerates ordinal subword alignments.
- Open Access
Effects of biological fixation methods on stimulated Raman scattering microscopy signal
Stimulated Raman Scattering (SRS) microscopy enables label-free imaging of cells and tissues in their native state with chemical specificity. However, there are often experimental advantages of chemical fixation of samples prior to imaging, which can introduce perturbations that may alter the native state of the samples, and possibly impact the SRS signal. In this study, we systematically characterize the effects of common fixatives (Paraformaldehyde, Formalin, Glutaraldehyde, Methanol, Ethanol) on the preservation of cellular integrity and molecular composition in a Hela cell model as observed by SRS microscopy. We demonstrate how the different fixatives can influence lipid and protein content, and overall cell morphology, with significant implications for the accuracy of quantitative SRS microscopy. Our findings indicate that Paraformaldehyde (PFA) imposes minimal disruption to cellular and molecular states compared to the other fixatives tested, and suggest Glutaraldehyde (GA) as a suitable alternative for SRS imaging. This study provides insights for the choice of the optimal sample preparation, enabling more reliable SRS-based studies for the evaluation of cellular processes and disease mechanisms where fixation is used.
Open AccessEffects of biological fixation methods on stimulated Raman scattering microscopy signal
Stimulated Raman Scattering (SRS) microscopy enables label-free imaging of cells and tissues in their native state with chemical specificity. However, there are often experimental advantages of chemical fixation of samples prior to imaging, which can introduce perturbations that may alter the native state of the samples, and possibly impact the SRS signal. In this study, we systematically characterize the effects of common fixatives (Paraformaldehyde, Formalin, Glutaraldehyde, Methanol, Ethanol) on the preservation of cellular integrity and molecular composition in a Hela cell model as observed by SRS microscopy. We demonstrate how the different fixatives can influence lipid and protein content, and overall cell morphology, with significant implications for the accuracy of quantitative SRS microscopy. Our findings indicate that Paraformaldehyde (PFA) imposes minimal disruption to cellular and molecular states compared to the other fixatives tested, and suggest Glutaraldehyde (GA) as a suitable alternative for SRS imaging. This study provides insights for the choice of the optimal sample preparation, enabling more reliable SRS-based studies for the evaluation of cellular processes and disease mechanisms where fixation is used.
- Open Access
Integrated analysis of multimodal long-read epigenetic assays
Long-read sequencing assays that detect base modifications are becoming increasingly important research tools for the study of epigenetic regulation, especially with the development of DiMeLo-seq and similar methods that deposit non-native base modifications to mark a range of epigenetic features such as protein-DNA interactions and chromatin accessibility. A main benefit of these methods is their inherent capacity for multimodality, enabling the encoding of multiple genomic signals onto single nucleic acid molecules. However, there are limited tools available for visualization and statistical analysis of this type of multimodal data. Here we introduce dimelo-toolkit , a python package built to enable flexible visualizations and easy integration into custom data processing workflows. We demonstrate the utility of dimelo-toolkit ’s preset visualizations of multiple base modifications in long-read single-molecule sequencing data with a novel extension of the DiMeLo-seq protocol that can capture three separate aspects of chromatin state on the same single reads: target protein binding, CpG methylation, and chromatin accessibility. We apply this multimodal method to simultaneously map chromatin accessibility, CpG methylation, and LMNB1 and CTCF binding patterns, respectively, in GM12878 cells. Our flexible design allows us to investigate previously unexplored technical biases that arise when working with this type of multimodal data. Additionally, we show that dimelo-toolkit enables analysis for a wide range of other long-read sequencing methods, such as mapping endogenous patterns in RNA base modifications with direct RNA sequencing. This software tool will pave the way for developing well-optimized protocols and help unlock previously inaccessible biological insights.
Open AccessIntegrated analysis of multimodal long-read epigenetic assays
Long-read sequencing assays that detect base modifications are becoming increasingly important research tools for the study of epigenetic regulation, especially with the development of DiMeLo-seq and similar methods that deposit non-native base modifications to mark a range of epigenetic features such as protein-DNA interactions and chromatin accessibility. A main benefit of these methods is their inherent capacity for multimodality, enabling the encoding of multiple genomic signals onto single nucleic acid molecules. However, there are limited tools available for visualization and statistical analysis of this type of multimodal data. Here we introduce dimelo-toolkit , a python package built to enable flexible visualizations and easy integration into custom data processing workflows. We demonstrate the utility of dimelo-toolkit ’s preset visualizations of multiple base modifications in long-read single-molecule sequencing data with a novel extension of the DiMeLo-seq protocol that can capture three separate aspects of chromatin state on the same single reads: target protein binding, CpG methylation, and chromatin accessibility. We apply this multimodal method to simultaneously map chromatin accessibility, CpG methylation, and LMNB1 and CTCF binding patterns, respectively, in GM12878 cells. Our flexible design allows us to investigate previously unexplored technical biases that arise when working with this type of multimodal data. Additionally, we show that dimelo-toolkit enables analysis for a wide range of other long-read sequencing methods, such as mapping endogenous patterns in RNA base modifications with direct RNA sequencing. This software tool will pave the way for developing well-optimized protocols and help unlock previously inaccessible biological insights.
- Open Access
Preselection CD4 + CD8 + thymocytes modulate TCR responsiveness following TCRβ selection
Modulation of TCR sensitivity during positive selection is critical to avoid negative selection and direct thymocytes into their appropriate lineage. Thymocytes just prior to positive selection (preselection) are highly responsive to low affinity self-ligands and are also actively rearranging their TCRα locus as they await a positive selection signal. Preselection DP thymocytes were thought to be relatively homogeneous and TCR modulation during this stage had not been previously described. Here we provide evidence for progressive gene expression changes within the preselection DP thymocyte population that correlates with a gradual loss of TCR responsiveness and a defect in upregulating TCR target genes associated with the CD4 fate. We relate these observations to the link between positive selection and T cell lineage commitment.
Open AccessPreselection CD4 + CD8 + thymocytes modulate TCR responsiveness following TCRβ selection
Modulation of TCR sensitivity during positive selection is critical to avoid negative selection and direct thymocytes into their appropriate lineage. Thymocytes just prior to positive selection (preselection) are highly responsive to low affinity self-ligands and are also actively rearranging their TCRα locus as they await a positive selection signal. Preselection DP thymocytes were thought to be relatively homogeneous and TCR modulation during this stage had not been previously described. Here we provide evidence for progressive gene expression changes within the preselection DP thymocyte population that correlates with a gradual loss of TCR responsiveness and a defect in upregulating TCR target genes associated with the CD4 fate. We relate these observations to the link between positive selection and T cell lineage commitment.
- Open Access
Single-Nucleus Analysis of Human White Adipose Tissue Reveals Adipocyte Subsets with Distinct Metabolic Profiles
Anatomic location of white adipose tissue is a determinant of cardiometabolic risk. To understand differences within/between adipose depots, we generated 65,668 single-nucleus transcriptomes from human subcutaneous or intraabdominal adipose tissue (SAT/IAT). Unsupervised analysis revealed 26 adipose-resident cell clusters including two subpopulations of mature adipocytes, characterized by high vs. low expression of adipocyte maturation genes (ADIPOMAT hi vs. ADIPOMAT lo ). ADIPOMAT lo adipocytes demonstrate a low-differentiation, pro-inflammatory, and pro-fibrotic transcriptome. IAT-resident ADIPOMAT lo were more abundant in higher BMI donors, while SAT-resident ADIPOMAT lo associated with impaired glycemia. TSHZ3 was identified as a candidate regulator of ADIPOMAT lo transcriptome. TSHZ3 knockdown in adipogenic progenitors inhibited differentiation, with downregulation of early adipogenic regulators (e.g. CEBPA/B, PPARG) and mature adipocyte genes. Heterozygous deletion of Tshz3 in mice reduced SAT and IAT weight. Here, we show that adipocyte subsets with distinct transcriptomic signature reside in human WAT; altered TSHZ3-mediated transcriptional regulation may contribute to low-maturation subpopulation linked to metabolic disease.
Open AccessSingle-Nucleus Analysis of Human White Adipose Tissue Reveals Adipocyte Subsets with Distinct Metabolic Profiles
Anatomic location of white adipose tissue is a determinant of cardiometabolic risk. To understand differences within/between adipose depots, we generated 65,668 single-nucleus transcriptomes from human subcutaneous or intraabdominal adipose tissue (SAT/IAT). Unsupervised analysis revealed 26 adipose-resident cell clusters including two subpopulations of mature adipocytes, characterized by high vs. low expression of adipocyte maturation genes (ADIPOMAT hi vs. ADIPOMAT lo ). ADIPOMAT lo adipocytes demonstrate a low-differentiation, pro-inflammatory, and pro-fibrotic transcriptome. IAT-resident ADIPOMAT lo were more abundant in higher BMI donors, while SAT-resident ADIPOMAT lo associated with impaired glycemia. TSHZ3 was identified as a candidate regulator of ADIPOMAT lo transcriptome. TSHZ3 knockdown in adipogenic progenitors inhibited differentiation, with downregulation of early adipogenic regulators (e.g. CEBPA/B, PPARG) and mature adipocyte genes. Heterozygous deletion of Tshz3 in mice reduced SAT and IAT weight. Here, we show that adipocyte subsets with distinct transcriptomic signature reside in human WAT; altered TSHZ3-mediated transcriptional regulation may contribute to low-maturation subpopulation linked to metabolic disease.
- Open Access
- 8 September 2025
- Published by Springer Nature in Nature Methods
- Vol. 22 (9) , 1-10
Scvi-hub: an actionable repository for model-driven single-cell analysis
The growing availability of single-cell omics datasets presents new opportunities for reuse, while challenges in data transfer, normalization and integration remain a barrier. Here we present scvi-hub: a platform for efficiently sharing and accessing single-cell omics datasets using pretrained probabilistic models. It enables immediate execution of fundamental tasks like visualization, imputation, annotation and deconvolution on new query datasets using state-of-the-art methods, with massively reduced storage and compute requirements. We show that pretrained models support efficient analysis of large references, including the CZI CELLxGENE Discover Census. Scvi-hub is built within the scvi-tools open-source environment and integrated into scverse. Scvi-hub offers a scalable and user-friendly framework for accessing and contributing to a growing ecosystem of ready-to-use models and datasets, thus putting the power of atlas-level analysis at the fingertips of a broad community of users.
- 8 September 2025
- Published by Springer Nature in Nature Methods
- Vol. 22 (9) , 1-10
Open AccessScvi-hub: an actionable repository for model-driven single-cell analysis
The growing availability of single-cell omics datasets presents new opportunities for reuse, while challenges in data transfer, normalization and integration remain a barrier. Here we present scvi-hub: a platform for efficiently sharing and accessing single-cell omics datasets using pretrained probabilistic models. It enables immediate execution of fundamental tasks like visualization, imputation, annotation and deconvolution on new query datasets using state-of-the-art methods, with massively reduced storage and compute requirements. We show that pretrained models support efficient analysis of large references, including the CZI CELLxGENE Discover Census. Scvi-hub is built within the scvi-tools open-source environment and integrated into scverse. Scvi-hub offers a scalable and user-friendly framework for accessing and contributing to a growing ecosystem of ready-to-use models and datasets, thus putting the power of atlas-level analysis at the fingertips of a broad community of users.
- Open Access
DiMeLo-cito: a one-tube protocol for mapping protein-DNA interactions reveals CTCF bookmarking in mitosis
Genome regulation relies on complex and dynamic interactions between DNA and proteins. Recently, powerful methods have emerged that leverage third-generation sequencing to map protein-DNA interactions genome-wide. For example, Directed Methylation with Long-read sequencing (DiMeLo-seq) enables mapping of protein-DNA interactions along long, single chromatin fibers, including in highly repetitive genomic regions. However, DiMeLo-seq involves lossy centrifugation-based wash steps that limit its applicability to many sample types. To address this, we developed DiMeLo-cito, a single-tube, wash-free protocol that maximizes the yield and quality of genomic DNA obtained for long-read sequencing. This protocol enables the interrogation of genome-wide protein binding with as few as 100,000 cells and without the requirement of a nuclear envelope, enabling confident measurement of protein-DNA interactions during mitosis. Using this protocol, we detected strong binding of CTCF to mitotic chromosomes in diploid human cells, in contrast with earlier studies in karyotypically unstable cancer cell lines, suggesting that CTCF “bookmarks” specific sites critical for maintaining genome architecture across cell divisions. By expanding the capabilities of DiMeLo-seq to a broader range of sample types, DiMeLo-cito can provide new insights into genome regulation and organization.
Open AccessDiMeLo-cito: a one-tube protocol for mapping protein-DNA interactions reveals CTCF bookmarking in mitosis
Genome regulation relies on complex and dynamic interactions between DNA and proteins. Recently, powerful methods have emerged that leverage third-generation sequencing to map protein-DNA interactions genome-wide. For example, Directed Methylation with Long-read sequencing (DiMeLo-seq) enables mapping of protein-DNA interactions along long, single chromatin fibers, including in highly repetitive genomic regions. However, DiMeLo-seq involves lossy centrifugation-based wash steps that limit its applicability to many sample types. To address this, we developed DiMeLo-cito, a single-tube, wash-free protocol that maximizes the yield and quality of genomic DNA obtained for long-read sequencing. This protocol enables the interrogation of genome-wide protein binding with as few as 100,000 cells and without the requirement of a nuclear envelope, enabling confident measurement of protein-DNA interactions during mitosis. Using this protocol, we detected strong binding of CTCF to mitotic chromosomes in diploid human cells, in contrast with earlier studies in karyotypically unstable cancer cell lines, suggesting that CTCF “bookmarks” specific sites critical for maintaining genome architecture across cell divisions. By expanding the capabilities of DiMeLo-seq to a broader range of sample types, DiMeLo-cito can provide new insights into genome regulation and organization.