Wei Xu

Wei Xu

Auto-generated profile, based on publicly available information

Wei Xu

Centre for Global Health and Molecular Epidemiology, Usher Institute, University of Edinburgh, Edinburgh, UK

Research Focus & Metrics

Academic Career Length

4Years (2023-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
2 January 2025 in Nature Communications
Verified
Journal Article
Open access
PDF file
WXWei Xu
Wei Xu
IMInes Mesa-Eguiagaray
Ines Mesa-Eguiagaray
DMDavid M. Morris
David M. Morris
CWChengjia Wang
Chengjia Wang
Calum David GrayCalum David Gray
Calum David Gray
WCWilliam P. Cawthorn
William P. Cawthorn
Bone marrow adipose tissue is a distinct adipose subtype comprising more than 10% of fat mass in healthy humans. However, the functions and pathophysiological correlates of this tissue are unclear, and its genetic determinants remain unknown. Here, we use deep learning to measure bone marrow adiposity in the femoral head, total hip, femoral diaphysis, and spine from MRI scans of approximately 47,000 UK Biobank participants, including over 41,000 white and over 6300 non-white participants. We then establish the heritability and genome-wide significant associations for bone marrow adiposity at each site. Our meta-GWAS in the white population finds 67, 147, 134, and 174 independent significant single nucleotide polymorphisms, which map to 54, 90, 43, and 100 genes for the femoral head, total hip, femoral diaphysis, and spine, respectively. Transcriptome-wide association studies, colocalization analyses, and sex-stratified meta-GWASes in the white participants further resolve functional and sex-specific genes associated with bone marrow adiposity at each site. Finally, we perform a multi-ancestry meta-GWAS to identify genes associated with bone marrow adiposity across the different bone regions and across ancestry groups. Our findings provide insights into BMAT formation and function and provide a basis to study the impact of BMAT on human health and disease.

12 September 2023 in eLife
Verified
Journal Article
Open access
RLReagan Lee
Reagan Lee
WXWei Xu
Wei Xu
MDMarshall Dozier
Marshall Dozier
RMRuth McQuillan
Ruth McQuillan
ETEvropi Theodoratou
Evropi Theodoratou
JFJonine Figueroa
Jonine Figueroa
COVID-19 has strained population breast mammography screening programs that aim to diagnose and treat breast cancers earlier. As the pandemic has affected countries differently, we aimed to quantify changes in breast screening volume and uptake during the first year of COVID-19 . We systematically searched Medline, the World Health Organization (WHO) COVID-19 database, and governmental databases. Studies covering January 2020 to March 2022 were included. We extracted and analyzed data regarding study methodology, screening volume, and uptake. To assess for risk of bias, we used the Joanna Briggs Institute (JBI) Critical Appraisal Tool. Twenty-six cross-sectional descriptive studies (focusing on 13 countries/nations) were included out of 935 independent records. Reductions in screening volume and uptake rates were observed among eight countries. Changes in screening participation volume in five nations with national population-based screening ranged from –13 to –31%. Among two countries with limited population-based programs, the decline ranged from –61 to –41%. Within the USA, population participation volumes varied ranging from +18 to –39%, with suggestion of differences by insurance status (HMO, Medicare, and low-income programs). Almost all studies had high risk of bias due to insufficient statistical analysis and confounding factors. The extent of COVID-19-induced reduction in breast screening participation volume differed by region and data suggested potential differences by healthcare setting (e.g., national health insurance vs. private healthcare). Recovery efforts should monitor access to screening and early diagnosis to determine whether prevention services need strengthening to increase the coverage of disadvantaged groups and reduce disparities.

23 September 2025 in Nature Communications
Verified
Journal Article
Open access
PDF file
WXWei Xu
Wei Xu
IMInes Mesa-Eguiagaray
Ines Mesa-Eguiagaray
DMDavid M. Morris
David M. Morris
CWChengjia Wang
Chengjia Wang
Calum David GrayCalum David Gray
Calum David Gray
WCWilliam P. Cawthorn
William P. Cawthorn
Bone marrow adiposity changes in diverse diseases, but the full scope of these, and whether they are directly influenced by marrow adiposity, remains unknown. To address this, we previously measured the bone marrow fat fraction of the femoral head, total hip, femoral diaphysis, and spine of over 48,000 UK Biobank participants. Here, we first use these data for PheWAS to identify diseases associated with marrow adiposity at each site. This reveals associations with 47 incident diseases across 12 disease categories, including osteoporosis, fracture, type 2 diabetes, cardiovascular diseases, cancers, and other conditions that burden public health worldwide. Intriguingly, type 2 diabetes associates positively with spine bone marrow adiposity but negatively with marrow adiposity at femoral sites. We then establish PRSs based on bone-marrow-fat-fraction-associated SNPs and use PRS-PheWAS and Mendelian randomization to explore causal associations between marrow adiposity and disease. PRS-PheWAS reveals that genetic predisposition to increased marrow adiposity is positively associated with osteoporosis and fractures. Mendelian randomization further suggests that increased marrow adiposity at the diaphysis and total hip is causally associated with osteoporosis. Our findings substantially advance understanding of how marrow adiposity impacts human health and highlight its potential as a biomarker and/or therapeutic target for diverse human diseases.

21 March 2025 in Journal of Global Health
Verified
Journal Article
Open access
XZXuan Zhou
Xuan Zhou
NYNan Yang
Nan Yang
WXWei Xu
Wei Xu
XLXue Li
Xue Li
ASAthina Spiliopoulou
Athina Spiliopoulou
ETEvropi Theodoratou
Evropi Theodoratou
Diabetes mellitus is a chronic condition characterised by high levels of blood glucose (hyperglycaemia). It arises when the body either fails to produce sufficient insulin or cannot effectively utilise the insulin it produces. It is estimated that diabetes mellitus affects approximately 537 million people worldwide, with a global prevalence of 10.5% in 2021, and the […]

10 March 2026 in Current Osteoporosis Reports
Verified
Journal Article
Open access
WXWei Xu
Wei Xu
CWChengjia Wang
Chengjia Wang
WCWilliam P Cawthorn
William P Cawthorn
Purpose of Review: Bone marrow adipose tissue (BMAT) is a significant fat depot with distinct skeletal, haematological and metabolic roles. It increases with ageing, osteoporosis, metabolic disease, and cancer, and is emerging as a biomarker for fracture risk. Quantification of bone marrow adiposity (BMA) has relied on MRI, proton spectroscopy, computed tomography, or histology, but large-scale studies have been limited by requiring labour-intensive analysis. Deep learning (DL) now enables scalable, automated BMA measurement, with greatest progress in MRI. This review highlights recent advances. Recent Findings: DL models, especially U-Nets, have been applied to UK Biobank MRI data, allowing site-specific BM fat fraction (BMFF) measurement or calvarial BMA estimation in tens of thousands of participants. These breakthroughs have revealed robust associations between BMA and age, sex, ethnicity, bone mineral density, adiposity, and metabolic traits, while uncovering site-specific patterns. Genome-wide association studies of BMFF and calvarial BMA have defined their genetic architecture, identifying hundreds of loci enriched for pathways in oestrogen signalling, adipogenesis, and skeletal remodelling. Phenome-wide association studies demonstrate links between altered BMFF and osteoporosis, fracture, type 2 diabetes, cardiovascular disease, and diverse other conditions, with Mendelian randomisation providing the first causal evidence that increased femoral BMFF contributes to osteoporosis. Despite successes, challenges remain, including extending analyses to non-European ancestries and validating DL pipelines in clinical settings. Summary: Collectively, DL-enabled BMA quantification has established BMAT as a clinically relevant, genetically tractable fat depot and provides new opportunities for mechanistic insight, risk prediction, and therapeutic targeting in musculoskeletal, metabolic, and other diseases.

6 October 2026 in SSRN Electronic Journal
Verified
Preprint
Open access
SSSamuel Sjöström
Samuel Sjöström
WXWei Xu
Wei Xu
DMDavid M. Morris
David M. Morris
CWChengjia Wang
Chengjia Wang
Calum David GrayCalum David Gray
Calum David Gray
WCWilliam P. Cawthorn
William P. Cawthorn
Background: Bone marrow adiposity has emerging links to type 2 diabetes (T2D), displaying opposite associations for axial verses appendicular skeletal sites. Thus, marrow adiposity distribution may outperform single-site measurements for T2D prediction. Methods: Using deep learning, we measured the bone marrow fat fraction (BMFF) in the spine, femoral head, total hip, and femoral diaphysis from 40,363 UK Biobank MRI participants. We then used principal component analysis to derive PC2, a multi-site measure that captures higher spine and lower femoral BMFF. We compared PC2 with regional BMFF using logistic and Cox models for incident T2D, including sex-stratified analyses and comparison with waist-to-hip ratio (WHR), an established T2D risk factor. Findings: PC2 correlated positively with spine BMFF and inversely with femoral BMFF. Among the marrow adiposity measures tested, PC2 showed the strongest association with T2D (OR 1.720 [1·631–1·815]; p<2×10-16). Phenome-wide analyses identified T2D as the strongest cardiometabolic association for PC2 and spine BMFF, with PC2 showing larger effects. PC2 did not significantly outperform WHR or spine BMFF as a standalone predictor; however, adding PC2 to an age- and sex-adjusted WHR model improved discrimination more than adding spine BMFF (ΔAUC 0·027 vs 0·007; DeLong p=0·069) and yielded stronger reclassification (NRI 0·297 [0·202–0·392] vs 0·154 [0·068–0·240]; IDI 0·004 [0·002–0·006] vs 0·001 [0.001–0·002]). Interpretation: PC2, a novel BMFF distribution parameter, is more strongly associated with T2D than single-site BMFF and may complement established risk factors for T2D prediction. Funding: Medical Research Council; British Heart Foundation; NHS Lothian; Cancer Research UK.