Anna Lloyd

Real‐World Evidence (RWE) is increasingly used in submissions to regulatory agencies and health technology assessment bodies (HTAbs) to support the efficacy and effectiveness of new medicines and indications. However, there is limited information on the RWE characteristics that impact its role in approval and reimbursement decisions. To investigate these characteristics, we developed FRAME: a Framework for Real‐world evidence Assessment to Mitigate Evidence uncertainties for efficacy/effectiveness. We compiled a list of medicinal product indications where RWE supported the efficacy of interventional trials or assessed effectiveness in observational settings. FRAME was applied to a prioritized subset of these submissions to authorities from North America, Europe, and Australia. For each product indication, we extracted information on characteristics describing the submission, clinical context, strength of evidence, and process factors from publicly available assessment reports. Of the 87 identified medicinal product indications, 15 were prioritized, covering 68 submissions and 76 RWE studies across 11 authorities in scope. Four main results emerged: (i) low granularity within assessment reports on the analyzed variables, limiting the learnings from analyzing them; (ii) variability in how RWE was assessed within and across regulatory agencies and HTAbs; (iii) a positive association between the proportion of positive comments from authorities on RWE studies and their impact on decision making. Particularly, a large effect size was consistently noted when RWE was considered primary evidence; and (iv) limited use of advanced RWE study designs. These findings support five recommendations to enhance shared learning on RWE, clarify its evidentiary value, and generate evidence to better support authorities' decision making.

A novel protein, PID-5, has been shown to be a requirement for germline immortality and has recently been implicated in RNA-induced epigenetic silencing in the Caenorhabditis elegans embryo. Importantly, it has been shown to contain both an eTudor and aminopeptidase P-related domain. However, the silencing mechanism has not yet been fully characterised. In this study, bioinformatic tools were used to compare pre-existing aminopeptidase P molecular structures to the AlphaFold2-predicted aminopeptidase P-related domain of PID-5 (PID-5 APP-RD). Structural homology, metal composition, inhibitor-bonding interactions, and the potential for dimerisation were critically assessed through computational techniques, including structural superimposition and protein-ligand docking. Results from this research suggest that the metallopeptidase-like domain shares high structural homology with known aminopeptidase P enzymes and possesses the canonical ‘pita-bread fold’. However, the absence of conserved metal-coordinating residues indicates that only a single Zn²⁺ may be bound at the active site. The PID-5 APP-RD may form transient interactions with a known aminopeptidase P inhibitor and may therefore recognise substrates in a comparable way to the known structures. However, loss of key catalytic residues suggests the domain will be inactive. Further evidence suggests that heterodimerisation with C. elegans aminopeptidase P is feasible and therefore PID-5 is predicted to regulate proteolytic cleavage in the silencing pathway. PID-5 may interact with PID-2 to bring aminopeptidase P activity to the Z-granule, where it could influence WAGO-4 activity to ensure the balanced production of 22G-RNA signals for transgenerational silencing. Targeted experiments into APPs implicated in malaria and cancer are required in order to build upon the biological and therapeutic significance of this research.

The Berkeley Open Infrastructure for Network Computing (BOINC) is an open source client-server middleware system created to allow projects with large computational requirements, usually set in the scientific domain, to utilize a technically unlimited number of volunteer machines distributed over large physical distances. However various problems exist deploying applications over these heterogeneous machines using BOINC: applications must be ported to each machine architecture type, the project server must be trusted to supply authentic applications, applications that do not regularly checkpoint may lose execution progress upon volunteer machine termination and applications that have dependencies may find it difficult to run under BOINC. To solve such problems we introduce virtual BOINC, or V-BOINC, where virtual machines are used to run computations on volunteer machines. Application developers can then compile their applications on a single architecture, check pointing issues are solved through virtualization API's and many security concerns are addressed via the virtual machine's sandbox environment. In this paper we focus on outlining a unique approach on how virtualization can be introduced into BOINC and demonstrate that V-BOINC offers acceptable computational performance when compared to regular BOINC. Finally we show that applications with dependencies can easily run under V-BOINC in turn increasing the computational potential volunteer computing offers to the general public and project developers.