Browsing by Author "Shroff, Raghav"

Now showing 1 - 2 of 2
  • Thumbnail Image
    Item
    Directed evolution of an orthogonal transcription engine for programmable gene expression in eukaryotes
    (Elsevier, 2025) Kar, Shaunak; Gardner, Elizabeth C.; Javanmardi, Kamyab; Boutz, Daniel R.; Shroff, Raghav; Horton, Andrew P.; Segall-Shapiro, Thomas H.; Ellington, Andrew D.; Gollihar, Jimmy; Bioengineering
    T7 RNA polymerase (RNAP) has enabled orthogonal control of gene expression and recombinant protein production across diverse prokaryotic host chassis organisms for decades. However, the absence of 5′ methyl guanosine caps on T7 RNAP-derived transcripts has severely limited its utility and widespread adoption in eukaryotic systems. To address this shortcoming, we evolved a fusion enzyme combining T7 RNAP with the single subunit capping enzyme from African swine fever virus using Saccharomyces cerevisiae. We isolated highly active variants of this fusion enzyme, which exhibited roughly two orders of magnitude higher protein expression compared to the wild-type enzyme. We demonstrate the programmable control of gene expression using T7 RNAP-based genetic circuits in yeast and validate enhanced performance of these engineered variants in mammalian cells. This study presents a robust, orthogonal gene regulatory system applicable across diverse eukaryotic hosts, enhancing the versatility and efficiency of synthetic biology applications.
  • Thumbnail Image
    Item
    DLPacker: Deep learning for prediction of amino acid side chain conformations in proteins
    (Wiley, 2022) Misiura, Mikita; Shroff, Raghav; Thyer, Ross; Kolomeisky, Anatoly B.; Center for Theoretical Biological Physics
    Prediction of side chain conformations of amino acids in proteins (also termed “packing”) is an important and challenging part of protein structure prediction with many interesting applications in protein design. A variety of methods for packing have been developed but more accurate ones are still needed. Machine learning (ML) methods have recently become a powerful tool for solving various problems in diverse areas of science, including structural biology. In this study, we evaluate the potential of deep neural networks (DNNs) for prediction of amino acid side chain conformations. We formulate the problem as image-to-image transformation and train a U-net style DNN to solve the problem. We show that our method outperforms other physics-based methods by a significant margin: reconstruction RMSDs for most amino acids are about 20% smaller compared to SCWRL4 and Rosetta Packer with RMSDs for bulky hydrophobic amino acids Phe, Tyr, and Trp being up to 50% smaller.