Biology subjects

Sanborn, A. L.

Publications and source records attributed to Sanborn, A. L..

4 recordsLinked to original sources

bioRxiv2023

ATOM-1: A Foundation Model for RNA Structure and Function Built on Chemical Mapping Data

RNA-based medicines and RNA-targeting drugs are emerging as promising new approaches for treating disease. Optimizing these therapeutics by naive experimental screening is a time-consuming and expensive process, while rational design requires an accurate understanding of the structure and function of RNA. To address this design challenge, we present ATOM-1, the first RNA foundation model trained on chemical mapping data, enabled by data collection strategies purposely developed for machine learning training. Using small probe neural networks on top of ATOM-1 embeddings, we demonstrate that this model has developed rich internal representations of RNA. Trained on limited amounts of additional data, these small networks achieve state-of-the-art accuracy on key RNA prediction tasks, suggesting that this approach can enable the design of therapies across the RNA landscape.

bioinformatics↗

bioRxiv2022

Mapping variation in the morphological landscape of human cells with optical pooled CRISPRi screening

The contributions of individual genes to cell-scale morphology and cytoskeletal organization are challenging to define due to the wide intercellular variation of these complex phenotypes. We leveraged the controlled nature of image-based pooled screening to assess the impact of CRISPRi knockdown of 366 genes on cell and nuclear morphology in human U2OS osteosarcoma cells. Screen scale-up was facilitated by a new, efficient barcode readout method that successfully genotyped 85% of cells. Phenotype analysis using a deep learning algorithm, the {beta}-variational autoencoder, produced a feature embedding space distinct from one derived from conventional morphological profiling, but detected similar gene hits while requiring minimal design decisions. We found 45 gene hits and visualized their effect by rationally constrained sampling of cells along the direction of phenotypic shift. By relating these phenotypic shifts to each other, we construct a quantitative and interpretable space of morphological variation in human cells.

cell biology↗

bioRxiv2020

Simple biochemical features underlie transcriptional activation domain diversity and dynamic, fuzzy binding to Mediator

Gene activator proteins comprise distinct DNA-binding and transcriptional activation domains (ADs). Because few ADs have been described, we tested domains tiling all yeast transcription factors for activation in vivo and identified 150 ADs. By mRNA display, we showed that 73% of ADs bound the Med15 subunit of Mediator, and that binding strength was correlated with activation. AD-Mediator interaction in vitro was unaffected by a large excess of free activator protein, pointing to a dynamic mechanism of interaction. Structural modeling showed that ADs interact with Med15 without shape complementarity ("fuzzy" binding). ADs shared no sequence motifs, but mutagenesis revealed biochemical and structural constraints. Finally, a neural network trained on AD sequences accurately predicted ADs in human proteins and in other yeast proteins, including chromosomal proteins and chromatin remodeling complexes. These findings solve the longstanding enigma of AD structure and function and provide a rationale for their role in biology.

molecular biology↗

bioRxiv2017

Cohesin Loss Eliminates All Loop Domains, Leading To Links Among Superenhancers And Downregulation Of Nearby Genes

The human genome folds to create thousands of intervals, called \"contact domains,\" that exhibit enhanced contact frequency within themselves. \"Loop domains\" form because of tethering between two loci - almost always bound by CTCF and cohesin - lying on the same chromosome. \"Compartment domains\" form when genomic intervals with similar histone marks co-segregate. Here, we explore the effects of degrading cohesin. All loop domains are eliminated, but neither compartment domains nor histone marks are affected. Loci in different compartments that had been in the same loop domain become more segregated. Loss of loop domains does not lead to widespread ectopic gene activation, but does affect a significant minority of active genes. In particular, cohesin loss causes superenhancers to co-localize, forming hundreds of links within and across chromosomes, and affecting the regulation of nearby genes. Cohesin restoration quickly reverses these effects, consistent with a model where loop extrusion is rapid.

genomics↗