Biology subjects

Lui, J.

Publications and source records attributed to Lui, J..

5 recordsLinked to original sources

bioRxiv2024

Transformer-Based Deep Learning Model with Latent Space Regularization for CRISPR-Cas Protein Sequence Classification

The discovery of the CRISPR-Cas system has significantly advanced genome editing, offering vast applications in medical treatments and life sciences research. Despite their immense potential, the existing CRISPR-Cas proteins still face challenges concerning size, delivery efficiency, and cleavage specificity. Addressing these challenges necessitates a deeper understanding of CRISPR-Cas proteins to enhance the design and discovery of novel Cas proteins for precision gene editing. In this study, we performed extensive deep-learning research on CRISPR-Cas proteins, aiming to develop a classification model capable of distinguishing CAS from non-CAS proteins, as well as discriminating sub-categories of CAS proteins, specifically CAS9 and CAS12. We developed two types of deep learning models: 1) a transformer encoder-based classification model, trained from scratch; and 2) a large protein language model fine-tuned on ProtBert, pre-trained on more than 200 million proteins. To boost learning efficiency for the model trained from scratch, we introduced a novel margin-based loss function to maximize inter-class separability and intra-class compactness in protein sequence embedding latent space of a transformer encoder. The experimental results show that the Fine-Tuned ProtBert-based (FTPB) classification model achieved accuracies of 99.06%, 94.42%, 96.80%, 97.57% for CAS9 vs. Non-CAS, CAS12 vs. Non-CAS, CAS9 vs. CAS12, and multi-class classification of CAS9 vs. CAS12 vs. Non-CAS, respectively. The Latent Space Regularized Max-Margin Transformer (LSRMT) model achieved classification accuracies of 99.81%, 99.81%, 99.06%, 99.27% for the same tasks, respectively. These results demonstrate the effectiveness of the proposed Max-Margin-based latent space regularization in enhancing model robustness and generalization capabilities. Remarkably, the LSRMT model, even when trained on a significantly smaller dataset, outperformed the fine-tuned state-of-the-art large protein model. The high classification accuracies achieved by the LSRMT model demonstrate its proficiency in identifying discriminative features of CAS proteins, marking a significant step towards advancing our understanding of CAS protein structures in future research endeavors.

bioinformatics↗

bioRxiv2020

Integrated multi-omics reveals common properties underlying stress granule and P-body formation

Non-membrane-bound compartments such as P-bodies (PBs) and stress granules (SGs) play important roles in the regulation of gene expression following environmental stresses. We have systematically determined the protein and mRNA composition of PBs and SGs formed in response to a common stress condition imposed by glucose depletion. We find that high molecular weight (HMW) complexes exist prior to glucose depletion that may act as seeds for the further condensation of proteins forming mature PBs and SGs. Both before and after glucose depletion, these HMW complexes are enriched for proteins containing low complexity and RNA binding domains. The mRNA content of these HMW complexes is enriched for long, structured mRNAs that become more poorly translated following glucose depletion. Many proteins and mRNAs are shared between PBs and SGs including several multivalent RNA binding proteins that may promote condensate interactions during liquid-liquid phase separation. Even where the precise identity of mRNAs and proteins localizing to PBs and SGs is distinct, the mRNAs and proteins share common biophysical and chemical features that likely trigger their phase separation.

biochemistry↗

bioRxiv2019

CoFe granules focus coordinated glycolytic mRNA localization and translation to fuel glucose fermentation

Glycolysis is a fundamental metabolic pathway for glucose catabolism across biology, and glycolytic enzymes are amongst the most abundant proteins in cells. Their expression at such levels provides a particular challenge. Here we demonstrate that the glycolytic mRNAs are localized to granules in yeast and human cells. Detailed live cell and smFISH studies in yeast show that the mRNAs are actively translated in granules, and this translation appears critical for the localization. Furthermore, this arrangement is likely to facilitate the higher level organisation and control of the glycolytic pathway. Indeed, the degree of fermentation required by cells is intrinsically connected to the extent of mRNA localization to granules. On this basis, we term these granules, Core Fermentation (CoFe) granules; they appear to represent translation factories allowing high-level co-ordinated enzyme synthesis for a critical metabolic pathway.

cell biology↗

bioRxiv2018

Translation factor mRNA granules direct protein synthetic capacity to regions of polarized growth

mRNA localization serves key functions in localized protein production making it critical that the translation machinery itself is present at these locations. Here we show that translation factor mRNAs are localized to distinct granules within yeast cells. In contrast to many mRNP granules, such as P-bodies and stress granules, which contain translationally repressed mRNAs, these granules harbor translated mRNAs under active growth conditions. The granules require Pablp for their integrity and are inherited by developing daughter cells in a She2p/ She3p dependent manner. These results point to a model where roughly half the mRNA for certain translation factors are specifically directed in granules toward the tip of the developing daughter cell where protein synthesis is most heavily required, which has particular implications for filamentous forms of growth. Such a feedforward mechanism would ensure adequate provision of the translation machinery where it is to be needed most over the coming growth cycle. SummaryThis study shows that mRNAs encoding a range of translation factors are localized to granules that get transported into the yeast daughter cell using the She2p/She3p machinery. This likely supports an intensification of protein synthetic activity to facilitate apical polarized growth.

cell biology↗

bioRxiv2018

Anti-α4β7 therapy targets lymphoid aggregates in the gastrointestinal tract of HIV-1 infected individuals

Herein, we present the first human study of anti-4{beta}7 therapy in a cohort of HIV-1 infected subjects with mild inflammatory bowel disease. 4{beta}7+ gut homing CD4+ T cells are early viral targets and contribute to HIV-1 pathogenesis, likely by seeding the gastrointestinal (GI) tract with HIV. Although, simianized anti-4{beta}7 monoclonal antibodies (Mab) have shown promise in preventing or attenuating the disease course of SIV in Non-Human Primate studies, the mechanisms of drug action remain elusive and the impact on HIV-1 persistence remains unanswered. By sampling the immune inductive and effector sites of the GI tract, we have discovered that anti-4{beta}7 therapy led to a significant and unexpected attenuation of lymphoid aggregates, most notably in the terminal ileum. Given that lymphoid aggregates serve as important sanctuary sites for establishing and maintaining viral reservoirs, their attrition by anti-4{beta}7 therapy has important implications for HIV-1 therapeutics and eradication efforts, and defines a rational basis for the continued evaluation of anti-4{beta}7 therapy in HIV-1 infection.\n\nOne Sentence SummaryAnti-4{beta}7 integrin therapy results in attrition of lymphoid aggregates within the gastrointestinal tract of HIV-1 infected individuals

immunology↗