Temporal inhibition of chromatin looping and enhancer accessibility during neuronal remodeling

During development, looping of an enhancer to a promoter is frequently observed in conjunction with temporal and tissue-specific transcriptional activation. The chromatin insulator-associated protein Alan Shepard (Shep) promotes Drosophila post-mitotic neuronal remodeling by repressing transcription of master developmental regulators, such as brain tumor (brat), specifically in maturing neurons. Since insulator proteins can promote looping, we hypothesized that Shep antagonizes brat promoter interaction with an as yet unidentified enhancer. Using chromatin conformation capture and reporter assays, we identified two enhancer regions that increase in looping frequency with the brat promoter specifically in pupal brains after Shep depletion. The brat promoters and enhancers function independently of Shep, ruling out direct repression of these elements. Moreover, ATAC-seq in isolated neurons demonstrates that Shep restricts chromatin accessibility of a key brat enhancer as well as other enhancers genome-wide in remodeling pupal but not larval neurons. These enhancers are enriched for chromatin targets of Shep and are located at Shep-inhibited genes, suggesting direct Shep inhibition of enhancer accessibility and gene expression during neuronal remodeling. Our results provide evidence for temporal regulation of chromatin looping and enhancer accessibility during neuronal maturation.

The leukemic oncogene EVI1 hijacks a MYC super-enhancer by CTCF-facilitated loops

Chromosomal rearrangements are a frequent cause of oncogene deregulation in human malignancies. Overexpression of EVI1 is found in a subgroup of acute myeloid leukemia (AML) with 3q26 chromosomal rearrangements, which is often therapy resistant. In AMLs harboring a t(3;8)(q26;q24), we observed the translocation of a MYC super-enhancer (MYC SE) to the EVI1 locus. We generated an in vitro model mimicking a patient-based t(3;8)(q26;q24) using CRISPR-Cas9 technology and demonstrated hyperactivation of EVI1 by the hijacked MYC SE. This MYC SE contains multiple enhancer modules, of which only one recruits transcription factors active in early hematopoiesis. This enhancer module is critical for EVI1 overexpression as well as enhancer-promoter interaction. Multiple CTCF binding regions in the MYC SE facilitate this enhancer-promoter interaction, which also involves a CTCF binding site upstream of the EVI1 promoter. We hypothesize that this CTCF site acts as an enhancer-docking site in t(3;8) AML. Genomic analyses of other 3q26-rearranged AML patient cells point to a common mechanism by which EVI1 uses this docking site to hijack enhancers active in early hematopoiesis.

Integrating convolution and self-attention improves language model of human genome for interpreting non-coding regions at base-resolution

Interpretation of non-coding genome remains an unsolved challenge in human genetics due to impracticality of exhaustively annotate biochemically active elements in all conditions. Deep learning based computational approaches emerge recently to help interpretating non-coding regions. Here we present LOGO (Language of Genome), a self-attention based contextualized pre-trained language model containing only 2 self-attention layers with 1 million parameters as a substantially light architecture that applies self-supervision techniques to learn bidirectional representations of unlabeled human reference genome. LOGO is then fine-tuned for sequence labelling task, and further extended to variant prioritization task via a special input encoding scheme of alternative alleles followed by adding a convolutional module. Experiments show that LOGO achieves 15% absolute improvement for promoter identification and up to 4.5% absolute improvement for enhancer-promoter interaction prediction. LOGO exhibits state-of-the-art multi-task predictive power on thousands of chromatin features with only 3% parameterization benchmarking against fully supervised model, DeepSEA and 1% parameterization against a recent BERT-based language model for human genome. For allelic-effect prediction, locality introduced by one dimensional convolution shows improved sensitivity and specificity for prioritizing non-coding variants associated with human diseases. In addition, we apply LOGO to interpret type 2 diabetes (T2D) GWAS signals and infer underlying regulatory mechanisms. We make a conceptual analogy between natural language and human genome and demonstrate LOGO is an accurate, fast, scalable, and robust framework to interpret non-coding regions for global sequence labeling as well as for variant prioritization at base-resolution.

Temporal inhibition of chromatin looping and enhancer accessibility during neuronal remodeling

ABSTRACTDuring development, looping of an enhancer to a promoter is frequently observed in conjunction with temporal and tissue-specific transcriptional activation. The chromatin insulator-associated protein Shep promotes Drosophila post-mitotic neuronal remodeling by repressing transcription of master developmental regulators, such as brain tumor (brat), specifically in maturing neurons. Since insulator proteins can promote looping, we hypothesized that Shep antagonizes brat promoter interaction with an as yet unidentified enhancer. Using chromatin conformation capture and reporter assays, we identified two novel enhancer regions that increase in looping frequency with the brat promoter specifically in pupal brains after Shep depletion. The brat promoters and enhancers function independently of Shep, ruling out direct repression of these elements. Moreover, ATAC-seq in isolated neurons demonstrated that Shep restricts chromatin accessibility of a key brat enhancer as well as other enhancers genome-wide in remodeling pupal but not larval neurons. These enhancers are enriched for chromatin targets of Shep and are located at Shep-inhibited genes, suggesting direct Shep inhibition of enhancer accessibility and gene expression during neuronal remodeling. Our results provide evidence for temporal regulation of chromatin looping and enhancer accessibility during neuronal maturation.

LOGO, a contextualized pre-trained language model of human genome flexibly adapts to various downstream tasks by fine-tuning

Interpretation of non-coding genome remains an unsolved challenge in human genetics due to impracticality of exhaustively annotate biochemically active elements in all conditions. Deep learning based computational approaches emerge recently to help interpretating non-coding regions. Here we present LOGO (Language of Genome), a self-attention based contextualized pre-trained language model that applies self-supervision techniques to learn bidirectional representations of unlabeled human reference genome and extend to a series of downstream tasks via fine-tuning. We also explore a novel knowledge embedded version of LOGO to incorporate prior human annotations. Experiments show that LOGO achieves 15% absolute improvement for promoter identification and up to 4.5% absolute improvement for enhancer-promoter interaction prediction. LOGO exhibits state-of-the-art predictive power on chromatin features with only 3% parameterization against fully supervised convolutional neural network, DeepSEA. Fine-tuned LOGO also shows outstanding performance in prioritizing non-coding variants associated with human diseases. In addition, we apply LOGO to interpret type 2 diabetes (T2D) GWAS signals and infer underlying regulatory mechanisms. We make a conceptual analogy between natural language and human genome and demonstrate LOGO is an accurate, fast, scalable, and robust framework with powerful adaptability to various tasks without substantial task-specific architecture modifications.

Differential regulation of alternative promoters emerges from unified kinetics of enhancer-promoter interaction

Many eukaryotic genes contain alternative promoters with distinct expression patterns. How these promoters are differentially regulated remains elusive. Here, we apply single-molecule imaging to quantify the transcriptional regulation of two alternative promoters (P1 and P2) of the Bicoid (Bcd) target gene hunchback in syncytial blastoderm Drosophila embryos. Contrary to the previous notion that Bcd only activates P2, we find that Bcd activates both promoters via the same two enhancers. P1 activation is less frequent and requires binding of more Bcd molecules than P2 activation. Using a theoretical model to relate promoter activity to enhancer states, we show that the two promoters follow common transcription kinetics driven by sequential Bcd binding at the two enhancers. Bcd binding at either enhancer primarily activates P2, while P1 activation relies more on Bcd binding at both enhancers. These results provide a quantitative framework for understanding the dynamics of complex eukaryotic gene regulation.

TAF4b transcription networks regulating early oocyte differentiation

Establishment of a healthy ovarian reserve is contingent upon numerous regulatory pathways during embryogenesis. Previously, mice lacking TBP-associated factor 4b (Taf4b) were shown to exhibit a diminished ovarian reserve. However, potential oocyte-intrinsic functions of TAF4b have not been examined. Here we use a combination of gene expression profiling and chromatin mapping to characterize the TAF4b gene regulatory network in mouse oocytes. We find that Taf4b-deficient oocytes display inappropriate expression of meiotic, chromatin, and X-linked genes, and unexpectedly we found a connection with Turner Syndrome pathways. Using Cleavage Under Targets and Release Using Nuclease (CUT&RUN), we observed TAF4b enrichment at genes involved in meiosis and DNA repair, some of which are differentially expressed in Taf4b-deficient oocytes. Interestingly, TAF4b target genes were enriched for Sp/KLF family motifs rather than TATA-box, suggesting an alternate mode of promoter interaction. Together, our data connects several gene regulatory nodes that contribute to the ovarian reserve.

Differential Bcd activation of two hunchback promoters emerges from unified kinetics of enhancer-promoter interaction

Many eukaryotic genes contain alternative promoters with distinct expression patterns. How these promoters are differentially regulated remains elusive. Here, we apply single-molecule imaging to quantify the transcriptional regulation of two alternative promoters (P1 and P2) of the Bicoid (Bcd) target gene hunchback in syncytial blastoderm Drosophila embryos. Contrary to the previous notion that Bcd only activates P2, we find that Bcd activates both promoters via the same two enhancers. P1 activation is less frequent and requires binding of more Bcd molecules than P2 activation. Using a theoretical model to relate promoter activity to enhancer states, we show that the two promoters follow common transcription kinetics driven by sequential Bcd binding at the two enhancers. Bcd binding at either enhancer primarily activates P2, while P1 activation relies more on Bcd binding at both enhancers. These results provide a quantitative framework for understanding the dynamics of complex eukaryotic gene regulation.

spatzie: An R package for identifying significant transcription factor motif co-enrichment from enhancer-promoter interactions

Genomic interactions provide important context to our understanding of the state of the genome. One question is whether specific transcription factor interactions give rise to genome organization. We introduce spatzie, an R package and a website that implements statistical tests for significant transcription factor motif cooperativity between enhancer-promoter interactions. We conducted controlled experiments under realistic simulated data from ChIP-seq to confirm spatzie is capable of discovering co-enriched motif interactions even in noisy conditions. We then use spatzie to investigate cell type specific transcription factor cooperativity within recent human ChIA-PET enhancer-promoter interaction data. The method is available online at https://spatzie.mit.edu.