Biology subjects

Dinger, M. E.

Publications and source records attributed to Dinger, M. E..

14 recordsLinked to original sources

bioRxiv2024

Chromatin remodeling with combined FACT and BET inhibition disrupts oncogenic transcription in Diffuse Midline Glioma

Aberrant epigenetic regulation is a hallmark of Diffuse Midline Glioma (DMG), an incurable pediatric brain tumor. The H3K27M driver histone mutation leads to transcriptional dysregulation, indicating that targeting the epigenome and transcription may be key therapeutic strategies against this highly aggressive cancer. One such target is the Facilitates Chromatin Transcription (FACT) histone chaperone. We found FACT to be enriched at developmental gene promoters, coinciding with regions of open chromatin and binding motifs of core DMG regulatory transcription factors. Furthermore, FACT interacted and co-localized with the Bromodomain and Extra-Terminal Domain (BET) protein BRD4 at promoters and enhancers, suggesting functional cooperation between FACT and BRD4 in DMG. In vitro, a combinatorial therapeutic approach using the FACT inhibitor CBL0137, coupled with BET inhibition revealed potent and synergistic cytotoxicity across a range of DMG cultures. These results were recapitulated in vivo, significantly extending survival in three independent orthotopic PDX models of DMG. Mechanistically, we show that CBL0137 treatment decreased chromatin accessibility, synergizing with BET inhibition to cause broad transcriptional collapse, silencing several key oncogenes including MYC, PDGFRA, MDM4 and SOX2, as well as causing alterations to the splicing landscape. Notably, this combination also elicited immune-related effects, including activation of the interferon response and antigen presentation mechanisms in DMG cells and induction of an activated state in macrophages and T cells, as demonstrated in an immunocompetent setting with spatial transcriptomics. Altogether, our data highlights the therapeutic promise of simultaneously targeting FACT and BET proteins in DMG, offering a dual tumor-intrinsic and immune-mediated strategy for combating this devastating pediatric brain tumor.

cancer biology↗

bioRxiv2024

Genes in Humans and Mice: Insights from Deep learning of 777K Bulk Transcriptomes

Mice are widely used as animal models in biomedical research, favored for their small size, ease of breeding, and anatomical and physiological similarities to humans1,2. However, discrepancies between mouse gene experimental results and the actual behavior of human genes are not uncommon, despite their shared DNA sequence similarity3-8. This suggests that DNA sequence similarity does not always reliably predict functional similarity. On the other hand, RNA-level gene expression could offer additional information about gene function9,10. In this study, we undertook characterization and inter-species comparison of human and mouse genes by applying innovative deep learning methodologies to a large dataset of 410K human and 366K mouse bulk RNA-seq samples. This was achieved by using gene representations from our Transformer-based GeneRAIN model11,12. These gene representations aggregate information from large gene expression datasets, and provide insights beyond DNA sequence similarity. We identified 2,407 human-mouse homologous genes with high DNA similarity but distinct RNA characteristics, and showed that these genes are more likely to have differing disease/phenotype associations between the two species. Additionally, we found 3,070 homologous genes with low similarity at both the DNA and RNA levels, suggesting the highest risk of discrepancies in study results between the two species. We propose that this approach will support future decision making around whether the mouse will be an appropriate model for studying specific human genes, and whether the results of specific mouse gene studies are likely to be recapitulated in humans. Our methodological innovations offer valuable lessons for future deep learning applications in cross-species omics data. The interspecies gene relationship findings from our study also contribute valuable insights into the gene biology and evolution of the two species.

bioinformatics↗

bioRxiv2024

Multifaceted Representation of Genes via Deep Learning of Gene Expression Networks

Accurate predictive modeling of human gene relationships would fundamentally transform our ability to uncover the molecular mechanisms that underpin key biological and disease processes. Recent studies have employed advanced AI techniques to model the complexities of gene networks using large gene expression datasets1-11. However, the extent and nature of the biological information these models can learn is not fully understood. Furthermore, the potential for improving model performance by using alternative data types, model architectures, and methodologies remains underexplored. Here, we developed GeneRAIN models by training on a large dataset of 410K human bulk RNA-seq samples, rather than single-cell RNA-seq datasets used by most previous studies. We showed that although the models were trained only on gene expression data, they learned a wide range of biological information well beyond gene expression. We introduced GeneRAIN-vec, a state-of-the-art, multifaceted vectorized representation of genes. Further, we demonstrated the capabilities and broad applicability of this approach by making 4,797 biological attribute predictions for each of 13,030 long non-coding RNAs (62.5 million predictions in total). These achievements stem from various methodological innovations, including experimenting with multiple model architectures and a new Binning-By-Gene normalization method. Comprehensive evaluation of our models clearly demonstrated that they significantly outperformed current state-of-the-art models3,12. This study improves our understanding of the capabilities of Transformer and self-supervised deep learning when applied to extensive expression data. Our methodological advancements offer crucial insights into refining these techniques. These innovations are set to significantly advance our understanding and exploration of biology.

bioinformatics↗

bioRxiv2023

Evolutionary conservation of embryonic DNA methylome remodelling in distantly related teleost species

Methylation of cytosines in the CG context (mCG) is the most abundant DNA modification in vertebrates that plays crucial roles in cellular differentiation and identity. After fertilization, DNA methylation patterns inherited from parental gametes are remodelled into a state compatible with embryogenesis. In mammals, this is achieved through the global erasure and re-establishment of DNA methylation patterns. However, in non-mammalian vertebrates like zebrafish, no global erasure has been observed. To investigate the evolutionary conservation and divergence of DNA methylation remodelling in teleosts, we generated base resolution DNA methylome datasets of developing medaka and medaka-zebrafish hybrid embryos. In contrast to previous reports, we show that medaka display comparable DNA methylome dynamics to zebrafish with high gametic mCG levels (sperm: [~]90%; egg: [~]75%), and adoption of a paternal-like methylome during early embryogenesis, with no signs of prior DNA methylation erasure. We also demonstrate that non-canonical DNA methylation (mCH) reprogramming at TGCT tandem repeats is a conserved feature of teleost embryogenesis. Lastly, we find remarkable evolutionary conservation of DNA methylation remodelling patterns in medaka-zebrafish hybrids, indicative of compatible DNA methylation maintenance machinery in far-related teleost species. Overall, these results suggest strong evolutionary conservation of DNA methylation remodelling pathways in teleosts, which is distinct from the global DNA methylome erasure and reestablishment observed in mammals.

developmental biology↗

bioRxiv2022

Human genomic DNA is widely interspersed with i-motif structures

DNA i-motif structures are formed in the nucleus of human cells and are believed to provide critical genomic regulation. While the existence of i-motif structures in human cells has been demonstrated by immunofluorescent staining and by characterisation of select model genes, the abundance and distribution of such structures in the human genome has remained unclear. Here we utilize high affinity i-motif immunoprecipitation followed by sequencing to map i-motifs in human genomic DNA. Validated by biolayer interferometry and circular dichroism spectroscopy, our approach identified over 650,000 i-motif structures in human genomic DNA. The i-motif structures are widely distributed throughout the human genome and are common among highly expressed genes and in genes upregulated in G0/G1 cell cycle phase. Our findings provide experimental evidence for the widespread formation of i-motif structures in human genomic DNA.

genetics↗

bioRxiv2022

Environmental and genetic disease modifiers of haploinsuffciency of A20

Monogenic diseases can often manifest diverse clinical phenotypes and cause diagnostic dilemmas. While monoallelic loss-of-function variants in TNFAIP3 (Haploinsufficiency of A20; HA20) cause a highly penetrant autoinflammatory disease, the variable expressivity suggest a role for additional genetic and environmental disease modifiers. Here, we identify critically ill children who inherited a family-specific TNFAIP3 deletion from one of their otherwise healthy parents. Each of the probands also inherited in trans a subtle loss-of-function I207L TNFAIP3 variant that is common in Oceania, originally introgressed from Denisovans. Modelling this compound heterozgous state in mice under specific pathogen free conditions demonstrated a reduced threshold to break immune tolerance. Exaggerated immune responses were precipitated by inheriting the two genetic hits on the TNFAIP3 checkpoint coupled with increasing the microbial challenge to immune tolerance, either by co-housing with pet store mice carrying a wild microbial burden or by transient dietary exposure to a chemical that diminishes the intestinal mucin barrier separating gut microbes from immune sensing systems. These data illuminate second-hit genetic and environmental modifiers contributing to complex inflammatory and autoimmune disease. Increased mechanistic understanding of the presence and contribution of disease modifiers will aid diagnostic and prognostic patient stratification and potentially reveal novel therapeutic opportunities.

immunology↗

bioRxiv2019

mity: A highly sensitive mitochondrial variant analysis pipeline for whole genome sequencing data

MotivationMitochondrial diseases (MDs) are the most common group of inherited metabolic disorders and are often challenging to diagnose due to extensive genotype-phenotype heterogeneity. MDs are caused by mutations in the nuclear or mitochondrial genome, where pathogenic mitochondrial variants are usually heteroplasmic and typically at much lower allelic fraction in the blood than affected tissues. Both genomes can now be readily analysed using unbiased whole genome sequencing (WGS), but most nuclear variant detection methods fail to detect low heteroplasmy variants in the mitochondrial genome. ResultsWe present mity, a bioinformatics pipeline for detecting and interpreting heteroplasmic SNVs and INDELs in the mitochondrial genome using WGS data. In 2,980 healthy controls, we observed on average 3,166x coverage in the mitochondrial genome using WGS from blood. mity utilises this high depth to detect pathogenic mitochondrial variants, even at low heteroplasmy. mity enables easy interpretation of mitochondrial variants and can be incorporated into existing diagnostic WGS pipelines. This could simplify the diagnostic pathway, avoid invasive tissue biopsies and increase the diagnostic rate for MDs and other conditions caused by impaired mitochondrial function. Availabilitymity is available from https://github.com/KCCG/mityunder an MIT license. Contactclare.puttick@crick.ac.uk, carolyn.sue@sydney.edu.au, MCowley@ccia.org.au

bioinformatics↗

bioRxiv2019

Phospho-tuning immunity through Denisovan, modern human, and mouse TNFAIP3 gene variants

Resisting or tolerating microbes are alternative strategies to survive infection, but little is known about the evolutionary mechanisms controlling this balance. Here, genomic analyses of anatomically modern humans, extinct Denisovan hominins, and mice revealed a series of missense variants in the immune response inhibitor A20 (encoded by TNFAIP3), substituting non-catalytic residues of the ubiquitin protease domain to diminish I{kappa}B-dependent phosphorylation and activation of A20. Two A20 variants with partial phosphorylation deficits appeared beneficial: one originating in Denisovans and introgressed in modern humans throughout Oceania, and another in a mouse strain resistant to Coxsackievirus. By contrast, a variant with 95% loss of phosphorylation caused spontaneous inflammatory disease in humans and mice. Analysis of the partial phosphorylation variant in mice revealed diminished tolerance of bacterial lipopolysaccharide or to poxvirus inoculation as trade-offs for enhanced immunity.\n\nOne Sentence SummaryModern and ancient variants reveal a genetically tunable element for balancing immunity and microbial tolerance.

immunology↗

bioRxiv2018

The Medical Genome Reference Bank: Whole genomes and phenotype of 2,570 healthy elderly

Population health research is increasingly focused on the genetic determinants of healthy ageing, but there is no public resource of whole genome sequences and phenotype data from healthy elderly individuals. Here we describe the Medical Genome Reference Bank (MGRB), comprising whole genome sequence and phenotype of 2,570 elderly Australians depleted for cancer, cardiovascular disease, and dementia. We analysed the MGRB for single-nucleotide, indel and structural variation in the nuclear and mitochondrial genomes. Individuals in the MGRB had fewer disease-associated common and rare germline variants, relative to both cancer cases and the gnomAD and UK BioBank cohorts, consistent with risk depletion. Pervasive age-related somatic changes were correlated with grip strength in men, suggesting blood-derived whole genomes may also provide a biologic measure of age-related functional deterioration. The MGRB provides a broadly applicable reference cohort for clinical genetics and genomic association studies, and for understanding the genetics of healthy ageing. This research has been conducted using the UK Biobank Resource under Application Number 17984.

genomics↗

bioRxiv2018

The Medical Genome Reference Bank: a whole-genome data resource of 4,000 healthy elderly individuals. Rationale and cohort design

Allele frequency data from human reference populations is of increasing value for filtering and assignment of pathogenicity to genetic variants. Aged and healthy populations are more likely to be selectively depleted of pathogenic alleles, and therefore particularly suitable as a reference populations for the major diseases of clinical and public health importance. However, reference studies of the healthy elderly have remained under-represented in human genetics. We have developed the Medical Genome Reference Bank (MGRB), a large-scale comprehensive whole-genome dataset of confirmed healthy elderly individuals, to provide a publicly accessible resource for health-related research, and for clinical genetics. It also represents a useful resource for studying the genetics of healthy aging. The MGRB comprises 4,000 healthy, older individuals with no reported history of cancer, cardiovascular disease or dementia, recruited from two Australian community-based cohorts. DNA derived from blood samples will be subject to whole genome sequencing. The MGRB will measure genome-wide genetic variation in 4,000 individuals, mostly of European decent, aged 60-95 years (mean age [≥] 75 years). The MGRB has committed to a policy of data sharing, employing a hierarchical data management system to maintain participant privacy and confidentiality, whilst maximizing research and clinical usage of the database. The MGRB will represent a dataset of international significance, broadly accessible to the clinical and genetic research community.

genomics↗

bioRxiv2018

Seave: a comprehensive web platform for storing and interrogating human genomic variation

Capability for genome sequencing and variant calling has increased dramatically, enabling large scale genomic interrogation of human disease. However, discovery is hindered by the current limitations in genomic interpretation, which remains a complicated and disjointed process. We introduce Seave, a web platform that enables variants to be easily filtered and annotated with in silico pathogenicity prediction scores and annotations from popular disease databases. Seave stores genomic variation of all types and sizes, and allows filtering for specific inheritance patterns, quality values, allele frequencies and gene lists. Seave is open source and deployable locally, or on a cloud computing provider, and works readily with gene panel, exome and whole genome data, scaling from single labs to multi-institution scale.

bioinformatics↗

bioRxiv2017

Universal Alternative Splicing Of Noncoding Exons

The human transcriptome is so large, diverse and dynamic that, even after a decade of investigation by RNA sequencing (RNA-Seq), we are yet to resolve its true dimensions. RNA-Seq suffers from an expression-dependent bias that impedes characterization of low-abundance transcripts. We performed targeted single-molecule and short-read RNA-Seq to survey the transcriptional landscape of a single human chromosome (Hsa21) at unprecedented resolution. Our analysis reaches the lower limits of the transcriptome, identifying a fundamental distinction between protein-coding and noncoding gene content: almost every noncoding exon undergoes alternative splicing, producing a seemingly limitless variety of isoforms. Analysis of syntenic regions of the mouse genome shows that few noncoding exons are shared between human and mouse, yet human splicing profiles are recapitulated on Hsa21 in mouse cells, indicative of regulation by a deeply conserved splicing code. We propose that noncoding exons are functionally modular, with alternative splicing generating an enormous repertoire of potential regulatory RNAs and a rich transcriptional reservoir for gene evolution.

genomics↗

bioRxiv2016

Machine-learning annotation of human splicing branchpoints

BackgroundThe branchpoint element is required for the first lariat-forming reaction in splicing. However due to difficulty in experimentally mapping at a genome-wide scale, current catalogues are incomplete.\n\nResultsWe have developed a machine-learning algorithm trained with empirical human branchpoint annotations to identify branchpoint elements from primary genome sequence alone. Using this approach, we can accurately locate branchpoints elements in 85% of introns in current gene annotations. Consistent with branchpoints as basal genetic elements, we find our annotation is unbiased towards gene type and expression levels. A major fraction of introns was found to encode multiple branchpoints raising the prospect that mutational redundancy is encoded in key genes. We also confirmed all deleterious branchpoint mutations annotated in clinical variant databases, and further identified thousands of clinical and common genetic variants with similar predicted effects.\n\nConclusionsWe propose the broad annotation of branchpoints constitutes a valuable resource for further investigations into the genetic encoding of splicing patterns, and interpreting the impact of common- and disease-causing human genetic variation on gene splicing.

bioinformatics↗

bioRxiv2016

High temporal resolution of gene expression dynamics in developing mouse embryonic stem cells

Investigations of transcriptional responses during developmental transitions typically use time courses with intervals that are not commensurate with the timescales of known biological processes. Moreover, such experiments typically focus on protein-coding transcripts, ignoring the important impact of long noncoding RNAs. We evaluated coding and noncoding expression dynamics at high temporal resolution (6-hourly) in differentiating mouse embryonic stem cells and report the effects of increased temporal resolution on the characterization of the underlying molecular processes. We present a refined resolution of global transcriptional alterations, including regulatory network interactions, coding and noncoding gene expression changes as well as alternative splicing events, many of which cannot be resolved by existing coarse developmental time-{-}-courses. We describe novel short lived and cycling patterns of gene expression and temporally dissect ordered gene expression at bidirectional promoters and responses to transcription factors. These findings demonstrate the importance of temporal resolution for understanding gene interactions in mammalian systems.\n\nLinks to dataData has been deposited into GEO: The Reviewer access link is: http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?token=cnglummejbkltyj@acc=GSE75028

systems biology↗