Machine learning framework to extract the biomarker potential of plasma IgG N-glycans towards disease risk stratification.

Effective management of chronic diseases and cancer can greatly benefit from disease-specific biomarkers that enable informative screening and timely diagnosis. IgG N-glycans found in human plasma have the potential to be minimally invasive disease-specific biomarkers for all stages of disease development due to their plasticity in response to various genetic and environmental stimuli. Data analysis and machine learning (ML) approaches can assist in harnessing the potential of IgG glycomics towards biomarker discovery and the development of reliable predictive tools for disease screening. This study proposes an ML-based N-glycomic analysis framework that can be employed to build, optimise, and evaluate multiple ML pipelines to stratify patients based on disease risk in an interpretable manner. To design and test this framework, a published colorectal cancer (CRC) dataset from the Study of Colorectal Cancer in Scotland (SOCCS) cohort (1999-2006) was used. In particular, among the different pipelines tested, an XGBoost-based ML pipeline, which was tuned using multi-objective optimisation, calibrated using an inductive Venn-Abers predictor (IVAP), and evaluated via a nested cross-validation (NCV) scheme, achieved a mean area under the Receiver Operating Characteristic Curve (AUC-ROC) of 0.771 when classifying between age-, and sex-matched healthy controls and CRC patients. This performance suggests the potential of using the relative abundance of IgG N-glycans to define populations at elevated CRC risk who merit investigation or surveillance. Finally, the IgG N-glycans that highly impact CRC classification decisions were identified using a global model-agnostic interpretability technique, namely Accumulated Local Effects (ALE). We envision that open-source computational frameworks, such as the one presented herein, will be useful in supporting the translation of glycan-based biomarkers into clinical applications.

Pandemic-response adenoviral vector and RNA vaccine manufacturing.

Rapid global COVID-19 pandemic response by mass vaccination is currently limited by the rate of vaccine manufacturing. This study presents a techno-economic feasibility assessment and comparison of three vaccine production platform technologies deployed during the COVID-19 pandemic: (1) adenovirus-vectored (AVV) vaccines, (2) messenger RNA (mRNA) vaccines, and (3) the newer self-amplifying RNA (saRNA) vaccines. Besides assessing the baseline performance of the production process, impact of key design and operational uncertainties on the productivity and cost performance of these vaccine platforms is quantified using variance-based global sensitivity analysis. Cost and resource requirement projections are computed for manufacturing multi-billion vaccine doses for covering the current global demand shortage and for providing annual booster immunisations. The model-based assessment provides key insights to policymakers and vaccine manufacturers for risk analysis, asset utilisation, directions for future technology improvements and future epidemic/pandemic preparedness, given the disease-agnostic nature of these vaccine production platforms.

Model Identifies Genetic Predisposition of Alzheimer's Disease as Key Decider in Cell Susceptibility to Stress.

Accumulation of unfolded/misfolded proteins in neuronal cells perturbs endoplasmic reticulum homeostasis, triggering a stress cascade called unfolded protein response (UPR), markers of which are upregulated in Alzheimer's disease (AD) brain specimens. We measured the UPR dynamic response in three human neuroblastoma cell lines overexpressing the wild-type and two familial AD (FAD)-associated mutant forms of amyloid precursor protein (APP), the Swedish and Swedish-Indiana mutations, using gene expression analysis. The results reveal a differential response to subsequent environmental stress depending on the genetic background, with cells overexpressing the Swedish variant of APP exhibiting the highest global response. We further developed a dynamic mathematical model of the UPR that describes the activation of the three branches of this stress response in response to unfolded protein accumulation. Model-based analysis of the experimental data suggests that the mutant cell lines experienced a higher protein load and subsequent magnitude of transcriptional activation compared to the cells overexpressing wild-type APP, pointing to higher susceptibility of mutation-carrying cells to stress. The model was then used to understand the effect of therapeutic agents salubrinal, lithium, and valproate on signalling through different UPR branches. This study proposes a novel integrated platform to support the development of therapeutics for AD.

An integrated biorefinery concept for conversion of sugar beet pulp into value-added chemicals and pharmaceutical intermediates.

Over 8 million tonnes of sugar beet are grown annually in the UK. Sugar beet pulp (SBP) is the main by-product of sugar beet processing which is currently dried and sold as a low value animal feed. SBP is a rich source of carbohydrates, mainly in the form of cellulose and pectin, including d-glucose (Glu), l-arabinose (Ara) and d-galacturonic acid (GalAc). This work describes the technical feasibility of an integrated biorefinery concept for the fractionation of SBP and conversion of these monosaccharides into value-added products. SBP fractionation is initially carried out by steam explosion under mild conditions to yield soluble pectin and insoluble cellulose fractions. The cellulose is readily hydrolysed by cellulases to release Glu that can then be fermented by a commercial yeast strain to produce bioethanol at a high yield. The pectin fraction can be either fully hydrolysed, using physico-chemical methods, or selectively hydrolysed, using cloned arabinases and galacturonases, to yield Ara-rich and GalAc-rich streams. These monomers can be separated using either Centrifugal Partition Chromatography (CPC) or ultrafiltration into streams suitable for subsequent enzymatic upgrading. Building on our previous experience with transketolase (TK) and transaminase (TAm) enzymes, the conversion of Ara and GalAc into higher value products was explored. In particular the conversion of Ara into l-gluco-heptulose (GluHep), that has potential therapeutic applications in hypoglycaemia and cancer, using a mutant TK is described. Preliminary studies with TAm also suggest GluHep can be selectively aminated to the corresponding chiral aminopolyol. The current work is addressing the upgrading of the remaining SBP monomer, GalAc, and the modelling of the biorefinery concept to enable economic and Life Cycle Analysis (LCA).

Comparative analysis of amino acid metabolism and transport in CHO variants with different levels of productivity.

Chinese hamster ovary (CHO) cells are widely used for the production of biopharmaceuticals; however, our understanding of several physiological elements that contribute to productivity is limited. One of these is amino acid transport and how its limitation and/or regulation might affect productivity. To further our understanding, we have examined the expression of 40 mammalian amino acid transporter genes during batch cultures of three CHO cell lines: a non-producer and two antibody-producing cell lines with different levels of productivity. In parallel, extracellular and intracellular levels of amino acids were quantified. The aim was to identify differences in gene regulation between cell lines and within culture. Our results show that three transporters associated with transport of taurine and ß-alanine, acidic amino acids and branched chain amino acids, are highly upregulated in both antibody-producing cell lines but not in the non-producer. Additionally, genes associated with the transport of amino acids related to the glutathione pathway (alanine, cysteine, cystine, glycine, glutamate) were found to be highly upregulated during the stationary phase of cell culture, correlating well with literature data on the importance of the pathway. Our analysis highlights potential markers for cell line selection and targets for process optimization.

Analysis of the landscape of biologically-derived pharmaceuticals in Europe: dominant production systems, molecule types on the rise and approval trends.

A thorough sort of the human drugs approved by the European Medicines Agency (EMA) between its establishment in 1995 until June 2012 is presented herein with a focus on biologically-derived pharmaceuticals. Over 200 (33%) of the 640 approved therapeutic drugs are derived from natural sources, produced via recombinant DNA technology, or generated through virus propagation. A breakdown based on production method, type of molecule and therapeutic category is presented. Current EMA approvals demonstrate that mammalian cells are the only choice for glycoprotein drugs, with Chinese hamster ovary cells being the dominant hosts for their production. On the other hand, bacterial cells and specifically Escherichia coli are the dominant hosts for protein-based drugs, followed by the yeast Saccharomyces cerevisiae. The latter is the dominant host for recombinant vaccine production, although egg-based production is still the main platform of vaccine provision. Our findings suggest that the majority of biologically-derived drugs are prescribed for cancer and related conditions, as well as the treatment of diabetes. The approval rate for biologically-derived drugs shows a strong upward trend for monoclonal antibodies and fusion proteins since 2009, while hormones, antibodies and growth factors remain the most populous categories. Despite a clear pathway for the approval of biosimilars set by the EMA and their potential to drive sales growth, we have only found approved biosimilars for three molecules. In 2012 there appears to be a slow-down in approvals, which coincides with a reported decline in the growth rate of biologics sales.