Biology subjects

Song, J. S.

Publications and source records attributed to Song, J. S..

18 recordsLinked to original sources

bioRxiv2026

Chromosome 3p deletion leads to extensive genomic alterations in diverse cancers and confers synthetic lethality in uveal melanoma

Chromosome 3p (chr3p) is frequently deleted in multiple cancers, indicating the presence of shared tumor suppressors. Analysis of genomic alterations in 33 different cancer types implicates the deletion or deleterious mutations of SET-domain-containing 2 (SETD2) at chr3p21 in significantly facilitating the formation of isochromosomes, consisting of two identical mirror-imaged arms, thereby promoting genomic instability conducive to large-scale chromosomal rearrangements and rapid cancer genome evolution. Fracturing of dicentric isochromosomes during cell division is pervasive and follows the dynamic fragmentation pattern of solids under impulse. Across cancers, isochromosomes form most frequently on chr8 to amplify the MYC-containing q-arm and chr17 to delete the TP53-containing p-arm. In the most aggressive uveal melanoma (UVM) subtype, chr3 deletion also includes MITF, a critical melanocyte differentiation and survival factor, and co-occurs with chr8q amplification. We demonstrate that MITF is a master transcriptional regulator of GNAQ/GNA11, mutated in 90% of UVM patients, and the associated synthetic-lethal genes identified by recent CRISPR screening studies. MITF maintains MAPK and calcium homeostasis in UVM, and its deletion is thus accidental, creating an early crisis during oncogenesis. We further show that MITF, MYC, and GNAQ/GNA11 form coupled regulatory feedback loops in the melanocyte lineage, and MITF deletion in UVM creates acute dependency on MYC-mediated rescue via chr8q amplification, often as a consequence of isochromosome formation. The discovered feedback loops predict both overall and relapse-free patient survival within the most aggressive UVM subtype, explain sensitivity to therapeutic gene perturbations, and inform effective combinatorial therapies.

cancer biology↗

bioRxiv2025

UV induces common cutaneous amyloid-like melanosomal protein aggregates

Misfolding of aggregation-prone proteins underpins diseases known as proteinopathies. One of these proteins, alpha-synuclein, is a component of aggregates in neurodegenerative conditions such as Parkinsons disease. The melanosomal protein PMEL, which forms physiologic amyloid scaffold structures on which melanin is organized in melanosomes, similarly ectopically accumulates in the dermis in many forms of cutaneous hyperpigmentation. Here, we demonstrate in a wide range of common clinical pigmentary disorders, as well as in primary melanocyte and mouse models examined by molecular, proteomic, and electron microscopic tools, that melanocytic alpha-synuclein is a prominent component of intracellular protein aggregates bound to similar proteins as in Parkinsons disease, as well as melanized extracellular protein deposits. Using the Real Time Quaking-Induced Conversion Assay (RT-QuIC), we demonstrate that UV induces misfolded melanosomal proteins to self-propagate, augmenting this pathology in prion-like fashion. CUT&RUN chromatin profiling and single-cell RNA-seq demonstrate that melanocytes utilize microphthalmia-associated transcription factor (MITF)-regulated autophagy to counteract protein aggregation, identifying aggregate removal as a core function of tanning. In contrast to extracellular aggregation, impaired intracellular aggregate removal contributes to melanocyte senescence, which conversely exacerbates chronic hypopigmentation and photoaging-related discoloration. These findings identify melanosomal proteinopathy as a common contributor to melanocyte dysfunction and suggest aggregate-focused management approaches.

molecular biology↗

bioRxiv2023

Chromatin structure and context-dependent sequence features control prime editing efficiency

Prime editor (PE) is a highly versatile CRISPR-Cas9 genome editing technique. The current constructs, however, have variable efficiency and may require laborious experimental optimization. This study presents statistical models for learning the salient epigenomic and sequence features of target sites modulating the editing efficiency and provides guidelines for designing optimal PEs. We found that both regional constitutive heterochromatin and local nucleosome occlusion of target sites impede editing, while position-specific G/C nucleotides in the primer binding site (PBS) and reverse transcription (RT) template regions of PE guide-RNA (pegRNA) yield high editing efficiency, especially for short PBS designs. The presence of G/C nucleotides was most critical immediately 5 to the protospacer adjacent motif (PAM) site for all designs. The effects of different last templated nucleotides were quantified and seen to depend on both PBS and RT template lengths. Our models found AGG to be the preferred PAM and detected a guanine nucleotide four bases downstream of PAM to facilitate editing, suggesting a hitherto-unrecognized interaction with Cas9. A neural network interpretation method based on nonextensive statistical mechanics further revealed multi-nucleotide preferences, indicating dependency among several bases across pegRNA. Our work clarifies previous conflicting observations and uncovers context-dependent features important for optimizing PE designs.

genomics↗

bioRxiv2022

Spectral clustering of single-cell multi-omics data on multilayer graphs

Single-cell sequencing technologies that simultaneously generate multimodal cellular profiles present opportunities for improved understanding of cell heterogeneity in tissues. How the multimodal information can be integrated to obtain a common cell type identification, however, poses a computational challenge. Multilayer graphs provide a natural representation of multi-omic single-cell sequencing datasets, and finding cell clusters may be understood as a multilayer graph partition problem. We introduce two spectral algorithms on multilayer graphs, spectral clustering on multilayer graphs (SCML) and the weighted locally linear (WLL) method, to cluster cells in multi-omic single-cell sequencing datasets. We connect these algorithms through a unifying mathematical framework that represents each layer using a Hamiltonian operator and a mixture of its eigenstates to integrate the multiple graph layers, demonstrating in the process that the WLL method is a rigorous multilayer spectral graph theoretic reformulation of the popular Seurat weighted nearest neighbor (WNN) algorithm. Implementing our algorithms and applying them to a CITE-seq dataset of cord blood mononuclear cells yields results similar to the Seurat WNN analysis. Our work thus extends spectral methods to multimodal single-cell data analysis. The code used in this study can be found at https://github.com/jssong-lab/sc-spectrum

bioinformatics↗

bioRxiv2021

Leveraging epigenomes and three-dimensional genome organization for interpreting regulatory variation

Understanding the impact of regulatory variants on complex phenotypes is a significant challenge because the genes and pathways that are targeted by such variants are typically unknown. Furthermore, a regulatory variant might influence a particular genes expression in a cell type or tissue-specific manner. Cell-type specific long-range regulatory interactions that occur between a distal regulatory sequence and a gene offers a powerful framework for understanding the impact of regulatory variants on complex phenotypes. However, high-resolution maps of such long-range interactions are available only for a handful of model cell lines. To address this challenge, we have developed L-HiC-Reg, a Random Forests based regression method to predict high- resolution contact counts in new cell lines, and a network-based framework to identify candidate cell line-specific gene networks targeted by a set of variants from a Genome-wide association study (GWAS). We applied our approach to predict interactions in 55 Roadmap Epigenome Consortium cell lines, which we used to interpret regulatory SNPs in the NHGRI GWAS catalogue. Using our approach, we performed an in-depth characterization of fifteen different phenotypes including Schizophrenia, Coronary Artery Disease (CAD) and Crohns disease. In CAD, we found differentially wired subnetworks consisting of known as well as novel gene targets of regulatory SNPs. Taken together, our compendium of interactions and associated network-based analysis pipeline offers a powerful resource to leverage long-range regulatory interactions to examine the context-specific impact of regulatory variation in complex phenotypes.

bioinformatics↗

bioRxiv2021

Predicting TCR-epitope Binding Specificity Using Deep Metric Learning and Multimodal Learning

Understanding the recognition of specific epitopes by cytotoxic T cells is a central problem in immunology. Although predicting binding between peptides and the class I Major Histocompatibility Complex (MHC) has had success, predicting interactions between T cell receptors (TCRs) and MHC class I-peptide complexes (pMHC) remains elusive. This paper utilizes a convolutional neural network model employing deep metric learning and multimodal learning to perform two critical tasks in TCR-epitope binding prediction: identifying the TCRs that bind a given epitope from a TCR repertoire, and identifying the binding epitope of a given TCR from a list of candidate epitopes. Our model can perform both tasks simultaneously and reveals that inconsistent preprocessing of CDR3B sequences can confound binding prediction. Applying a neural network interpretation method identifies key amino acid sequence patterns and positions within the TCR important for binding specificity. Contrary to the common assumption, known crystal structures of TCR-pMHC complexes show that the predicted salient amino acid positions are not necessarily the closest to the epitopes, implying that physical proximity may not be a good proxy for importance in determining TCR-epitope specificity. Our work thus provides insight into the learned predictive features of TCR-epitope binding specificity and advances associated classification tasks.

bioinformatics↗

bioRxiv2020

Measuring DNA mechanics on the genome scale

Mechanical deformations of DNA such as bending are ubiquitous and implicated in diverse cellular functions1. However, the lack of high-throughput tools to directly measure the mechanical properties of DNA limits our understanding of whether and how DNA sequences modulate DNA mechanics and associated chromatin transactions genome-wide. We developed an assay called loop-seq to measure the intrinsic cyclizability of DNA - a proxy for DNA bendability - in high throughput. We measured the intrinsic cyclizabilities of 270,806 50 bp DNA fragments that span the entire length of S. cerevisiae chromosome V and other genomic regions, and also include random sequences. We discovered sequence-encoded regions of unusually low bendability upstream of Transcription Start Sites (TSSs). These regions disfavor the sharp DNA bending required for nucleosome formation and are co-centric with known Nucleosome Depleted Regions (NDRs). We show biochemically that low bendability of linker DNA located about 40 bp away from a nucleosome edge inhibits nucleosome sliding into the linker by the chromatin remodeler INO80. The observation explains how INO80 can create promoter-proximal nucleosomal arrays in the absence of any other factors2 by reading the DNA mechanical landscape. We show that chromosome wide, nucleosomes are characterized by high DNA bendability near dyads and low bendability near the linkers. This contrast increases for nucleosomes deeper into gene bodies, suggesting that DNA mechanics plays a previously unappreciated role in organizing nucleosomes far from the TSS, where nucleosome remodelers predominate. Importantly, random substitution of synonymous codons does not preserve this contrast, suggesting that the evolution of codon choice has been impacted by selective pressure to preserve sequence-encoded mechanical modulations along genes. We also provide evidence that transcription through the TSS-proximal nucleosomes is impacted by local DNA mechanics. Overall, this first genome-scale map of DNA mechanics hints at a mechanical code with broad functional implications.

biophysics↗

bioRxiv2020

Riemannian geometry and statistical modeling correct for batch effects and control false discoveries in single-cell surface protein count data from CITE-seq

Recent advances in next generation sequencing-based single-cell technologies have allowed high-throughput quantitative detection of cell-surface proteins along with the transcriptome in individual cells, extending our understanding of the heterogeneity of cell populations in diverse tissues that are in different diseased states or under different experimental conditions. Count data of surface proteins from the cellular indexing of transcriptomes and epitopes by sequencing (CITE-seq) technology pose new computational challenges, and there is currently a dearth of rigorous mathematical tools for analyzing the data. This work utilizes concepts and ideas from Riemannian geometry to remove batch effects between samples and develops a statistical framework for distinguishing positive signals from background noise. The strengths of these approaches are demonstrated on two independent CITE-seq data sets in mouse and human. Python source code implementing the algorithms is available at https://github.com/jssong-lab/SAGACITE.

bioinformatics↗

bioRxiv2020

Epigenetic engineering of yeast reveals dynamic molecular adaptation to methylation stress and genetic modulators of specific DNMT3 family members

Cytosine methylation is a ubiquitous modification in mammalian DNA generated and maintained by several DNA methyltransferases (DNMTs) with partially overlapping functions and genomic targets. To systematically dissect the factors specifying each DNMTs activity, we engineered combinatorial knock-in of human DNMT genes in Komagataella phaffii, a yeast species lacking endogenous DNA methylation. Time-course expression measurements captured dynamic network-level adaptation of cells to DNMT3B1-induced DNA methylation stress and showed that coordinately modulating the availability of S-adenosyl methionine (SAM), the essential metabolite for DNMT-catalyzed methylation, is an evolutionarily conserved epigenetic stress response, also implicated in several human diseases. Convolutional neural networks trained on genome-wide CpG-methylation data learned distinct sequence preferences of DNMT3 family members. A simulated annealing interpretation method resolved these preferences into individual flanking nucleotides and periodic poly(A) tracts that rotationally position highly methylated cytosines relative to phased nucleosomes. Furthermore, the nucleosome repeat length defined the spatial unit of methylation spreading. Gene methylation patterns were similar to those in mammals, and hypo- and hypermethylation were predictive of increased and decreased transcription relative to control, respectively, in the absence of mammalian readers of DNA methylation. Introducing controlled epigenetic perturbations in yeast thus enabled characterization of fundamental genomic features directing specific DNMT3 proteins.

genomics↗

bioRxiv2019

Chd1 regulates repair of promoter-proximal DNA breaks to sustain hypertranscription in embryonic stem cells

Stem and progenitor cells undergo a global elevation of nascent transcription, or hypertranscription, during key developmental transitions involving rapid cell proliferation. The chromatin remodeler Chd1 binds to genes transcribed by RNA Polymerase (Pol) I and II and is required for hypertranscription in embryonic stem (ES) cells in vitro and the early post-implantation epiblast in vivo. Biochemically, Chd1 has been shown to facilitate transcription at least in part by removing nucleosomal barriers to elongation, but its mechanism of action in stem cells remains poorly understood. Here we report a novel role for Chd1 in the repair of promoter-proximal endogenous double-stranded DNA breaks (DSBs) in ES cells. An unbiased proteomics approach revealed that Chd1 interacts with several DNA repair factors including Atm, Parp1, Kap1 and Topoisomerase 2{beta}. We show that wild-type ES cells display high levels of phosphorylated H2A.X and Kap1 at chromatin, notably at rDNA in the nucleolus, in a Chd1-dependent manner. Loss of Chd1 leads to an extensive accumulation of DSBs at Chd1-bound Pol II-transcribed genes and rDNA. Genes prone to DNA breaks in Chd1 KO ES cells tend to be longer genes with GC-rich promoters, a more labile nucleosomal structure and roles in chromatin regulation, transcription and signaling. These results reveal a vulnerability of hypertranscribing stem cells to endogenous DNA breaks, with important implications for developmental and cancer biology.

molecular biology↗

bioRxiv2019

Knowledge-guided analysis of ‘omics’ data using the KnowEnG cloud platform

We present KnowEnG, a free-to-use computational system for analysis of genomics data sets, designed to accelerate biomedical discovery. It includes tools for popular bioinformatics tasks such as gene prioritization, sample clustering, gene set analysis and expression signature analysis. The system offers knowledge-guided data-mining and machine learning algorithms, where user-provided data are analyzed in light of prior information about genes, aggregated from numerous knowledge-bases and encoded in a massive Knowledge Network. KnowEnG adheres to FAIR principles: its tools are easily portable to diverse computing environments, run on the cloud for scalable and cost-effective execution of compute-intensive and data-intensive algorithms, and are interoperable with other computing platforms. They are made available through multiple access modes including a web-portal, and include specialized visualization modules. We present use cases and re-analysis of published cancer data sets using KnowEnG tools and demonstrate its potential value in democratization of advanced tools for the modern genomics era.

cancer biology↗

bioRxiv2018

Can cancer GWAS variants modulate immune cells in the tumor microenvironment?

Genome-wide association studies (GWAS) have hitherto identified several genetic variants associated with cancer susceptibility, but the molecular functions of these risk modulators remain largely uncharacterized. Recent studies have begun to uncover the regulatory potential of non-coding GWAS SNPs by using epigenetic information in corresponding cancer cell types and matched normal tissues. However, this approach does not explore the potential effect of risk germline variants on other important cell types that constitute the microenvironment of tumor or its precursor. This paper presents evidence that the breast cancer-associated variant rs3903072 may regulate the expression of CTSW in tumor infiltrating lymphocytes. CTSW is a candidate tumor-suppressor gene, with expression highly specific to immune cells and also positively correlated with breast cancer patient survival. Integrative analyses suggest a putative causative variant in a GWAS-linked enhancer in lymphocytes that loops to the 3 end of CTSW through three-dimensional chromatin interaction. Our work thus poses the possibility that a cancer-associated genetic variant might regulate a gene not only in the cell of cancer origin, but also in immune cells in the microenvironment, thereby modulating the immune surveillance by T lymphocytes and natural killer cells and affecting the clearing of early cancer initiating cells.

cancer biology↗

bioRxiv2018

SequencEnG: an Interactive Knowledge Base of Sequencing Techniques

Next-generation sequencing (NGS) techniques are revolutionizing biomedical research by providing powerful methods for generating genomic and epigenomic profiles. The rapid progress is posing an acute challenge to students and researchers to stay acquainted with the numerous available methods. We have developed an interactive online educational resource called SequencEnG (acronym for Sequencing Techniques Engine for Genomics) to provide a tree-structured knowledge base of 66 different sequencing techniques and step-by-step NGS data analysis pipelines comparing popular tools. SequencEnG is designed to facilitate barrier-free learning of current NGS techniques and provides a user-friendly interface for searching through experimental and analysis methods. SequencEnG is part of the project KnowEnG (Knowledge Engine for Genomics) and is freely available at http://education.knoweng.org/sequenceng/.

bioinformatics↗

bioRxiv2017

A unified computational framework for modeling genome-wide nucleosome landscape

Nucleosomes form the fundamental building blocks of eukaryotic chromatin, and previous attempts to understand the principles governing their genome-wide distribution have spurred much interest and debate in biology. In particular, the precise role of DNA sequence in shaping local chromatin structure has been controversial. This paper rigorously quantifies of the contribution of hitherto-debated sequence features - including G+C content, 10.5-bp periodicity, and poly(dA:dT) tracts - to three distinct aspects of genome-wide nucleosome landscape: occupancy, translational positioning and rotational positioning. Our computational framework simultaneously learns nucleosome number and nucleosome-positioning energy from genome-wide nucleosome maps. In contrast to other previous studies, our model can predict both in-vitro and in-vivo nucleosome maps in S. cerevisiae. We find that although G+C content is the primary determinant of MNase-derived nucleosome occupancy, MNase digestion biases may substantially influence this GC dependence. By contrast, poly(dA:dT) tracts are seen to deter nucleosome formation, regardless of the experimental method used. We further show that the 10.5-bp nucleotide periodicity facilitates rotational but not translational positioning. Applying our method to in-vivo nucleosome maps demonstrates that, for a subset of genes, the regularly-spaced nucleosome arrays observed around transcription start sites can be partially recapitulated by DNA sequence alone. Finally, in-vivo nucleosome occupancy derived from MNase-seq experiments around transcription termination sites can be mostly explained by the genomic sequence. Implications of these results and potential extensions of the proposed computational framework are discussed

bioinformatics↗

bioRxiv2017

High accuracy label-free classification of kinetic cell states from holographic cytometry

Digital holographic microscopy permits live and label-free visualization of adherent cells. Here we report the application of this approach for high accuracy kinetic quantitative cytometry. We identify twenty-six label-free optical and morphological features that are biologically independent. When used as a basis for machine learning, these features allow blind single cell classification with up to 95% accuracy. We present methods to control for inherent holographic noise, thereby establishing a set of reliable quantitative features. Together, these contributions permit continuous digital holographic cytometry for three or more days. Applying our approach to human melanoma cells treated with a panel of cancer therapeutics, we can track the response of each cell, simultaneously classifying multiple behaviors such as cell cycle length, motility, apoptosis, senescence, and heterogeneity of response to each therapeutic. Importantly, we demonstrate relationships between these phenotypes over time. This work thus provides an experimental and computational roadmap for low cost live-cell imaging and kinetic classification of heterogeneous adherent cell populations.

cancer biology↗

bioRxiv2017

ClusterEnG: An interactive educational web resource for clustering big data

SummaryClustering is one of the most common techniques used in data analysis to discover hidden structures by grouping together data points that are similar in some measure into clusters. Although there are many programs available for performing clustering, a single web resource that provides both state-of-the-art clustering methods and interactive visualizations is lacking. ClusterEnG (acronym for Clustering Engine for Genomics) provides an interface for clustering big data and interactive visualizations including 3D views, cluster selection and zoom features. ClusterEnG also aims at educating the user about the similarities and differences between various clustering algorithms and provides clustering tutorials that demonstrate potential pitfalls of each algorithm. The web resource will be particularly useful to scientists who are not conversant with computing but want to understand the structure of their data in an intuitive manner.\n\nAvailabilityClusterEnG is part of a bigger project called KnowEnG (Knowledge Engine for Genomics) and is available at http://education.knoweng.org/clustereng.\n\nContactsongi@illinois.edu

bioinformatics↗

bioRxiv2017

Emergent community agglomeration from data set geometry

In the statistical learning language, samples are snapshots of random vectors drawn from some unknown distribution. Such vectors usually reside in a high-dimensional Euclidean space, and thus, the \"curse of dimensionality\" often undermines the power of learning methods, including community detection and clustering algorithms, that rely on Euclidean geometry. This paper presents the idea of effective dissimilarity transformation (EDT) on empirical dissimilarity hyperspheres and studies its effects using synthetic and gene expression data sets. Iterating the EDT turns a static data distribution into a dynamical process purely driven by the empirical data set geometry and adaptively ameliorates the curse of dimensionality, partly through changing the topology of a Euclidean feature space [R]n into a compact hypersphere Sn. The EDT often improves the performance of hierarchical clustering via the automatic grouping information emerging from global interactions of data points. The EDT is not restricted to hierarchical clustering, and other learning methods based on pairwise dissimilarity should also benefit from the many desirable properties of EDT.\n\nPACS numbers: 89.20.Ff, 87.85.mg

bioinformatics↗

bioRxiv2017

Maximum Entropy Methods for Extracting the Learned Features of Deep Neural Networks

New architectures of multilayer artificial neural networks and new methods for training them are rapidly revolutionizing the application of machine learning in diverse fields, including business, social science, physical sciences, and biology. Interpreting deep neural networks, however, currently remains elusive, and a critical challenge lies in understanding which meaningful features a network is actually learning. We present a general method for interpreting deep neural networks and extracting network-learned features from input data. We describe our algorithm in the context of biological sequence analysis. Our approach, based on ideas from statistical physics, samples from the maximum entropy distribution over possible sequences, anchored at an input sequence and subject to constraints implied by the empirical function learned by a network. Using our framework, we demonstrate that local transcription factor binding motifs can be identified from a network trained on ChIP-seq data and that nucleosome positioning signals are indeed learned by a network trained on chemical cleavage nucleosome maps. Imposing a further constraint on the maximum entropy distribution also allows us to probe whether a network is learning global sequence features, such as the high GC content in nucleosome-rich regions. This work thus provides valuable mathematical tools for interpreting and extracting learned features from feed-forward neural networks.

bioinformatics↗