5683099 research outputs found
Sort by
Importance of sample size on the quality and utility of AI-based prediction models for healthcare
No full textRigorous study design and analytical standards are required to generate reliable findings in healthcare from artificial intelligence (AI) research. One crucial but often overlooked aspect is the determination of appropriate sample sizes for studies developing AI-based prediction models for individual diagnosis or prognosis. Specifically, the number of participants and outcome events required in datasets for model training and evaluation remains inadequately addressed. Most AI studies do not provide a rationale for their chosen sample sizes and frequently rely on datasets that are inadequate for training or evaluating a clinical prediction model. Among the ten principles of Good Machine Learning Practice established by the US Food and Drug Administration, the UK Medicines and Healthcare products Regulatory Agency, and Health Canada, guidance on sample size is directly relevant to at least three principles. To reinforce this recommendation, we outline seven reasons why inadequate sample size negatively affects model training, evaluation, and performance. Using a range of examples, we illustrate these issues and discuss the potentially harmful consequences for patient care and clinical adoption. Additionally, we address challenges associated with increasing sample sizes in AI research and highlight existing approaches and software for calculating the minimum sample sizes required for model training and evaluation
'The Crick’s Data Archive: Supporting Best Practice in Research Data Integrity' - Poster displayed at Crick Spring Technology Meeting, Tuesday 29th April 2025
No full textPoster documenting the research data archiving process to support research integrity at the Crick, inclusive of QR code to a training course examining best practice in research data sharing, for further reference.</p
The urgent need for newer drugs in routine HIV treatment in Africa: the case of Ghana
No full textAntiretroviral therapy (ART) has tremendously improved the quality of life of people living with HIV (PLWH). Through rigorous scientific research and development, newer, more effective, and less toxic antiretrovirals (ARVs) have been developed and are available to PLWH in high-income countries (HICs). Although Africa accounts for more than two-thirds of the global burden of HIV/AIDS, this large population does not readily have access to these newer and more effective ARVs. In some instances, new ARVs become available to PLWH in Africa over a decade after they have been approved for use by the Food and Drug Authorities (FDAs) in HICs. Since 2010, 35 new drug entities have been approved; of those, only 3 are in common use in Ghana and most of Sub-Saharan Africa. To achieve the 2030 goal of ending HIV/AIDS as a global health epidemic, it is critical to ensure equity in access to newer and effective ARVs across all regions, including Africa, where the majority of PLWH reside. We highlight here the urgent need to make newer ARVs available in Africa to ensure the realization of the Global End AIDS by 2030 goal
An automated workflow to address proteome complexity and the large search space problem in proteomics and HLA-I immunopeptidomics.
No full textAntigenic noncanonical epitope and novel protein discovery are research areas with therapeutical applications, predominantly done via mass spectrometry. The latter should rely on a well-characterized proteogenomic search space. Its size is barely known for antigenic noncanonical peptides and novel proteins, and this could impact on their identification. To address these issues, we here develop an automated workflow comprised of Sequoia for the creation of RNA sequencing informed and exhaustive sequence search spaces for various noncanonical peptide origins, and SPIsnake for pre-filtering and exploration of sequence search space prior to mass spectrometry searches. We apply our workflow to characterize the exact sizes of tryptic and nonspecific peptide sequence search spaces in a variety of definitions, their reduction when using RNA expression, their inflation by post-translational modifications, and the frequency of peptide sequence multimapping to different noncanonical origins. Furthermore, we explore the application of Sequoia and SPIsnake on HLA-I immunopeptidomes, thereby rescuing sensitivity in peptide identification when confronted with inflated search spaces. Taken together, Sequoia and SPIsnake pave the way for an educated development of methods addressing large-scale exhaustive proteogenomic discovery by exposing the consequences of database size inflation and ambiguity of peptide and protein sequence identification
Heterogeneity of thymic output in the elderly and its association with sex and smoking.
No full textBACKGROUND: Thymic involution with age leads to reduced T cell output and impaired adaptive immunity. However, the extent to which thymic activity persists later in life and how this contributes to immunological ageing remains unclear. This study aimed to assess the presence and function of thymic tissue in older adults and identify factors influencing residual thymopoiesis. METHODS: Patients aged ≥ 50 undergoing cardiothoracic surgery were recruited. Thymic structures within mediastinal adipose tissue were evaluated using histology, immunofluorescence, flow cytometry, TCR sequencing, and RNA sequencing. Recent thymic emigrants (RTEs) were quantified in peripheral blood and correlated with transcriptomic, epigenetic, and TCR repertoire data. Primary outcomes included thymic tissue identification, RTE frequency, and immune correlates. RESULTS: Functional thymic tissue was identified in mediastinal adipose tissue of older individuals. The frequency of CD31+CD4+ T cells (RTEs) positively correlated with the presence of thymic tissue. Thymic output showed substantial heterogeneity and was influenced by sex and smoking history. Thymic activity was associated with increased TCR repertoire diversity, improved immune protection to infections, and reduced epigenetic ageing. Detailed profiling uncovered functional and phenotypic heterogeneity within naïve CD4+ T cell subsets shaped by thymic activity. CONCLUSION: This study demonstrates that thymic function can persist into later life and is modulated by factors such as sex and smoking. These findings suggest that thymic activity during ageing is heterogenous and influenced by more than chronological age alone, with potential implications for immune competence in older adults
MRTF-dependent cytoskeletal dynamics drive efficient cell cycle progression.
No full textSerum response factor (SRF) and its cofactors, Myocardin-related transcription factors A/B (MRTF-A/B), regulate transcription of numerous cytoskeletal structural and regulatory genes, and most MRTF/SRF inactivation phenotypes reflect deficits in cytoskeletal dynamics. We show that MRTF-SRF activity is required for effective proliferation of both primary and immortalised fibroblast and epithelial cells. Cells lacking the MRTFs or SRF proliferate very slowly, express elevated levels of SASP factors and SA-b-galactosidase activity, and inhibit proliferation of co-cultured primary wildtype cells. They exhibit decreased levels of CDK1 and CKS2 proteins, and elevated levels of CDK inhibitors, usually CDKN1B/p27. These phenotypes, which can be fully reversed by re-expression of MRTF-A, are also seen in wildtype cells arrested by serum deprivation. Moreover, in wildtype cells direct interference with cytoskeletal dynamics through inhibition of ROCKs or myosin ATPase induces a similar proliferative defect to that seen in MRTF-null cells. MRTF-null cells exhibit multiple cytoskeletal defects, and markedly reduced contractility. We propose that MRTF-SRF signalling will be required for cell proliferation in cell types and environments where physical progression through cell cycle transitions requires high contractility
S-TIR-ring up TLR7 and TLR9: signaling domain substitutions clarify the TLR paradox
No full textIn systemic lupus erythematosus (SLE), autoimmunity often develops toward self nucleic acids. The nucleic acid receptors TLR7 and TLR9, which sense RNA and DNA, respectively, are critical for the generation of pathogenic autoimmune antibodies. Despite similarities in their downstream signaling cascades, these receptors play opposing roles in most mouse lupus models: TLR7 promotes disease, while TLR9 provides protection — an observation often referred to as “the TLR paradox.” To understand the basis of this dichotomy, Leibler et al. created genetically edited lupus-prone mice in which TLR7 receptors express the TLR intracellular signaling domain (TIR) that corresponds to TLR9, or vice versa. Their results revealed that the TIR domains contribute to the receptors’ opposing roles in SLE, shedding light into the TLR paradox in autoimmunity
EVOLVE-HBV mental health protocol
No full textThe overall prevalence of mental health conditions (depression, anxiety, quality of life and a poor global functioning) in persons living with chronic HBV infection has been reported to be higher than in the absence of Hepatitis B. At the beginning of our project, the EVOLVE team engaged in community dialogue to gain an appreciation of the HBV landscape within the uMkhanyakude district. We found low awareness of HBV among the community and identified a high need to increase education and awareness efforts. The community also identified a need to tackle stigma and we set out to recognise the potential impact of HBV infection on overall wellbeing, including mental health. We also recognise that there is a risk that a new diagnosis of HBV infection could cause anxiety and distress in some individuals. A specific objective of the study is to collect pilot data for mental health assessment to create a more holistic understanding of the multifaceted nature of HBV.</p
Mitochondrial ABHD11 inhibition drives sterol metabolism to modulate T-cell effector function.
No full textα/β-hydrolase domain-containing protein 11 (ABHD11) is a mitochondrial hydrolase that maintains the catalytic function of α-ketoglutarate dehydrogenase (α-KGDH), and its expression in CD4 + T-cells has been linked to remission status in rheumatoid arthritis (RA). However, the importance of ABHD11 in regulating T-cell metabolism and function is yet to be explored. Here, we show that pharmacological inhibition of ABHD11 dampens cytokine production by human and mouse T-cells. Mechanistically, the anti-inflammatory effects of ABHD11 inhibition are attributed to increased 24,25-epoxycholesterol (24,25-EC) biosynthesis and subsequent liver X receptor (LXR) activation, which arise from a compromised TCA cycle. The impaired cytokine profile established by ABHD11 inhibition is extended to two patient cohorts of autoimmunity. Importantly, using murine models of accelerated type 1 diabetes (T1D), we show that targeting ABHD11 suppresses cytokine production in antigen-specific T-cells and delays the onset of diabetes in vivo in female mice. Collectively, our work provides pre-clinical evidence that ABHD11 is an encouraging drug target in T-cell-mediated inflammation