Biology subjects

Chen, X. S.

Publications and source records attributed to Chen, X. S..

8 recordsLinked to original sources

bioRxiv2026

Cell-of-origin shapes tumor subtype in BRCA1-associated triple-negative breast cancer

Pathogenic germline BRCA1 variants greatly increase lifetime breast cancer risk, with no targeted prevention. Accrual of mutations following DNA repair errors in luminal progenitors is considered the chief mechanism of triple-negative breast cancer (TNBC) initiation, which cannot explain the organ specificity nor the phenotypic spectrum of the disease. Here we show that TNBC also originates in basal-myoepithelial cells, and that the cell of origin shapes TNBC subtype through aberrant epithelial differentiation, likely due to pleiotropic role of BRCA1 in regulating hormone signaling. Single-cell RNA sequencing of 67,001 cells from Wap-Cre Brca1/p53 mice and reanalysis of published Blg-Cre data show that aberrant basal-myoepithelial differentiation (Wap-Cre) generates the mesenchymal subtype (TNBC-M) and aberrant luminal differentiation (Blg-Cre) the basal-like subtype (TNBC-BL1). We also show that both subtypes occur at comparable frequency in BRCA1 carriers. Aberrant basal-myoepithelial differentiation precedes tumor formation, and only tumors derived from transplanted basal cells reproduce the basal-containing phenotype of spontaneous tumors. Comparison with a developmental atlas identifies de-differentiation of myoepithelial cells into basal cells lacking myoepithelial markers, a fetal state, whereas experimental metastases shift toward luminal states while retaining a TNBC-M-like transcriptional program. These findings support the inclusion of basal-myoepithelial and luminal cells in targeted prevention strategies for BRCA1 carriers, and adoption of TNBC subtyping for patient stratification and improving outcomes.

cancer biology↗

bioRxiv2026

Multivariate Random Forests for Cross-Modal Multi-Omics Integration

Multi-omics studies are widely used across many areas of biomedical research. In many diseases, some signals are shared across data types, while others are strongest in a single omics layer. Current multi-omics clustering methods often either merge all data types into a single representation, which can blur biology that is strong in one layer, or rely on linear structure that may miss more complex relationships across data types. We introduce O_SCPLOWMULTIC_SCPLOWRF, a random-forest-based method that handles complex data types and separates shared and modality-specific structure for multi-omics data. O_SCPLOWMULTIC_SCPLOWRF learns sample similarities across omics layers from multivariate random forests, combines them across data types, and uses the resulting weights to estimate the part of each omics layer that is predictable from the others. The remaining residual is treated as modality-specific signal, allowing shared and modality-specific similarities to be clustered separately. In simulations, O_SCPLOWMULTIC_SCPLOWRF recovered shared clusters as well as or better than established integrative methods while more reliably separating modality-specific signal under nonlinear data structures. In TCGA head and neck squamous cell carcinoma, the shared component aligned with the main subtype structure across established reference classifications, while gene- and miRNA-specific components revealed additional immune and developmental biology. In the ADNI cohort with matched blood DNA methylation and structural MRI, the shared cross-modal aging signal was associated with future conversion to mild cognitive impairment or Alzheimers disease, and a DNAm-specific residual signal showed exploratory additional information. These results show that O_SCPLOWMULTIC_SCPLOWRF can recover a common disease axis while retaining biologically meaningful signals specific to one data type. O_SCPLOWMULTIC_SCPLOWRF is available as an open-source R package at https://github.com/novawz/multiRF.

bioinformatics↗

bioRxiv2024

Structural Basis for Pol theta-Helicase DNA Binding and Microhomology-Mediated End-Joining

DNA double-strand breaks (DSBs) present a critical threat to genomic integrity, often precipitating genomic instability and oncogenesis. Repair of DSBs predominantly occurs through homologous recombination (HR) and non-homologous end joining (NHEJ). In HR-deficient cells, DNA polymerase theta (Pol{theta}) becomes critical for DSB repair via microhomology-mediated end joining (MMEJ), also termed theta-mediated end joining (TMEJ). Thus, Pol{theta} is synthetically lethal with BRCA1/2 and other HR factors, underscoring its potential as a therapeutic target in HR-deficient cancers. However, the molecular mechanisms governing Pol{theta}-mediated MMEJ remain poorly understood. Here we present a series of cryo-electron microscopy structures of the Pol{theta} helicase domain (Pol{theta}-hel) in complex with DNA containing 3'-overhang. The structures reveal the sequential conformations adopted by Pol{theta}-hel during the critical phases of DNA binding, microhomology searching, and microhomology annealing. The stepwise conformational changes within the Pol{theta}-hel subdomains and its functional dimeric state are pivotal for aligning the 3'-overhangs, facilitating the microhomology search and subsequent annealing necessary for DSB repair via MMEJ. Our findings illustrate the essential molecular switches within Pol{theta}-hel that orchestrate the MMEJ process in DSB repair, laying the groundwork for the development of targeted therapies against the Pol{theta}-hel.

molecular biology↗

bioRxiv2024

Molecular mechanism for regulating APOBEC3G DNA editing function by the non-catalytic domain

APOBEC3G (A3G) belongs to the AID/APOBEC cytidine deaminase family and is essential for antiviral immunity. It contains two zinc-coordinated cytidine-deaminase (CD) domains. The N-terminal CD1 domain is non-catalytic but has a strong affinity for nucleic acids, whereas the C-terminal CD2 domain catalyzes C-to-U editing in single-stranded DNA. The interplay between the two domains in DNA binding and editing is not fully understood. Here, our studies on rhesus macaque A3G (rA3G) show that the DNA editing function in linear and hairpin loop DNA is greatly enhanced by AA or GA dinucleotide motifs present downstream (in the 3-direction) but not upstream (in the 5-direction) of the target-C editing sites. The effective distance between AA/GA and the target-C sites depends on the local DNA secondary structure. We present two co-crystal structures of rA3G bound to ssDNA containing AA and GA, revealing the contribution of the non-catalytic CD1 domain in capturing AA/GA DNA and explaining our biochemical observations. Our structural and biochemical findings elucidate the molecular mechanism underlying the cooperative function between the non-catalytic and the catalytic domains of A3G, which is critical for its antiviral role and its contribution to genome mutations in cancer.

biochemistry↗

bioRxiv2024

PNAbind: Structure-based prediction of protein-nucleic acid binding using graph neural networks

The recognition and binding of nucleic acids (NAs) by proteins depends upon complementary chemical, electrostatic and geometric properties of the protein-NA binding interface. Structural models of protein-NA complexes provide insights into these properties but are scarce relative to models of unbound proteins. We present a deep learning approach for predicting protein-NA binding given the apo structure of a protein (PNAbind). Our method utilizes graph neural networks to encode spatial distributions of physicochemical and geometric properties of the protein molecular surface that are predictive of NA binding. Using global physicochemical encodings, our models predict the overall binding function of a protein and can discriminate between specificity for DNA or RNA binding. We show that such predictions made on protein structures modeled with AlphaFold2 can be used to gain mechanistic understanding of chemical and structural features that determine NA recognition. Using local encodings, our models predict the location of NA binding sites at the level of individual binding residues. Binding site predictions were validated against benchmark datasets, achieving AUROC scores in the range of 0.92-0.95. We applied our models to the HIV-1 restriction factor APOBEC3G and show that our predictions are consistent with experimental RNA binding data.

bioinformatics↗

bioRxiv2023

Structural basis for polyuridine tract recognition by SARS-CoV-2 Nsp15

SARS-CoV-2 non-structural protein 15 (Nsp15) is critical for productive viral replication and evasion of host immunity. The uridine-specific endoribonuclease activity of Nsp15 mediates the cleavage of the polyuridine [poly(U)] tract of the negative-strand coronavirus genome to minimize the formation of dsRNA that activates the host antiviral interferon signaling. However, the molecular basis for the recognition and cleavage of the poly(U) tract by Nsp15 is incompletely understood. Here, we present cryogenic electron microscopy (cryoEM) structures of SARS-CoV-2 Nsp15 bound to viral replication intermediate dsRNA containing poly(U) tract at 2.7-3.3 [A] resolution. The structures reveal one copy of dsRNA binds to the sidewall of an Nsp15 homohexamer, spanning three subunits in two distinct binding states. The target uracil is dislodged from the base-pairing of the dsRNA by amino acid residues W332 and M330 of Nsp15, and the dislodged base is entrapped at the endonuclease active site center. Up to 20 A/U base pairs are anchored on the Nsp15 hexamer, which explains the basis for a substantially shortened poly(U) sequence in the negative strand coronavirus genome compared to the long poly(A) tail in its positive strand. Our results provide mechanistic insights into the unique immune evasion strategy employed by coronavirus Nsp15.

molecular biology↗

bioRxiv2022

An Automatized Workflow to Study Mechanistic Indicators for Driver Gene Prediction with Moonlight

Prediction of tumor suppressors and oncogenes, also called driver genes, is an essential step in understanding cancer development and discovering potential novel treatments. We recently proposed Moonlight as a bioinformatics framework to predict driver genes and analyze them in a system-biology-oriented manner based on -omics integration. Moonlight uses gene expression as a primary data source and combines it with patterns related to cancer hallmarks and regulatory networks to identify oncogenic mediators. Once the oncogenic mediators are identified, it is important to include extra levels of evidence, called mechanistic indicators, to identify driver genes and to link the observed changes in gene expression to the underlying alteration that promotes them. Such a mechanistic indicator could be for example a mutation in the regulatory regions for the candidate gene or mutations in the regulator itself. In this work, we developed new functionalities and release Moonlight2, to provide the user with the mutation-based mechanistic indicator to streamline the analyses of this second layer of evidence. The function analyzes mutation information in a cancer cohort to classify them into driver and passenger mutations. Moreover, the function estimates the potential effect of a mutation on the transcriptional, translational, or protein structure/function level. Those oncogenic mediators with at least one driver mutation are retained as the final set of driver genes. We applied Moonlight2 and the newly developed function to a case study on Basal-like breast cancer subtype using data from The Cancer Genome Atlas. We found six oncogenes (SF3B4, EBNA1BP2, KRTCAP2, ZBTB8OS, RUNX2, and POLR2J) and ten tumor suppressor genes (KIF26B, NR5A2, ARHGAP25, EMCN, ARL15, PCOLCE, TPK1, TEK, KIR2DL4, and GMFG) containing a driver mutation in their promoter region, possibly explaining their deregulation. The Moonlight2R source code is available at https://github.com/ELELAB/Moonlight2R.

bioinformatics↗

bioRxiv2018

Hexameric helicase G40P unwinds DNA in single base pair steps

Most replicative helicases are hexameric, ring-shaped motor proteins that translocate on and unwind DNA. Despite extensive biochemical and structural investigations, how their translocation activity is utilized chemo-mechanically in DNA unwinding is poorly understood. We examined DNA unwinding by G40P, a DnaB-family helicase, using a single-molecule fluorescence assay with a single base pair resolution. The high-resolution assay revealed that G40P by itself is a very weak helicase that stalls at barriers as small as a single GC base pair and unwinds DNA with the step size of a single base pair. Single ATP{gamma}S binding could stall unwinding, demonstrating highly coordinated ATP hydrolysis between the six identical subunits. We observed frequent slippage of the helicase, which is fully suppressed by the primase DnaG. We anticipate that these findings allow a better understanding on the fine balance of thermal fluctuation activation and energy derived from hydrolysis.

biophysics↗