Deep learning connects DNA traces to transcription to reveal predictive features beyond enhancer-promoter contact.

Chromatin architecture plays an important role in gene regulation. Recent advances in super-resolution microscopy have made it possible to measure chromatin 3D structure and transcription in thousands of single cells. However, leveraging these complex data sets with a computationally unbiased method has been challenging. Here, we present a deep learning-based approach to better understand to what degree chromatin structure relates to transcriptional state of individual cells. Furthermore, we explore methods to "unpack the black box" to determine in an unbiased manner which structural features of chromatin regulation are most important for gene expression state. We apply this approach to an Optical Reconstruction of Chromatin Architecture dataset of the Bithorax gene cluster in Drosophila and show it outperforms previous contact-focused methods in predicting expression state from 3D structure. We find the structural information is distributed across the domain, overlapping and extending beyond domains identified by prior genetic analyses. Individual enhancer-promoter interactions are a minor contributor to predictions of activity.

Thymic tuft cells promote an IL-4-enriched medulla and shape thymocyte development.

The thymus is responsible for generating a diverse yet self-tolerant pool of T cells1. Although the thymic medulla consists mostly of developing and mature AIRE+ epithelial cells, recent evidence has suggested that there is far greater heterogeneity among medullary thymic epithelial cells than was previously thought2. Here we describe in detail an epithelial subset that is remarkably similar to peripheral tuft cells that are found at mucosal barriers3. Similar to the periphery, thymic tuft cells express the canonical taste transduction pathway and IL-25. However, they are unique in their spatial association with cornified aggregates, ability to present antigens and expression of a broad diversity of taste receptors. Some thymic tuft cells pass through an Aire-expressing stage and depend on a known AIRE-binding partner, HIPK2, for their development. Notably, the taste chemosensory protein TRPM5 is required for their thymic function through which they support the development and polarization of thymic invariant natural killer T cells and act to establish a medullary microenvironment that is enriched in the type 2 cytokine, IL-4. These findings indicate that there is a compartmentalized medullary environment in which differentiation of a minor and highly specialized epithelial subset has a non-redundant role in shaping thymic function.