Login / Signup

Associations of Genetically Predicted Vitamin B 12 Status across the Phenome.

Marie-Joe DibKourosh R AhmadiLoukas ZagkosDipender GillBrooke MorrisPaul ElliottAbbas DehghanJoanna Tzoulaki
Published in: Nutrients (2022)
Variation in vitamin B 12 levels has been associated with a range of diseases across the life-course, the causal nature of which remains elusive. We aimed to interrogate genetically predicted vitamin B 12 status in relation to a plethora of clinical outcomes available in the UK Biobank. Genome-wide association study (GWAS) summary data obtained from a Danish and Icelandic cohort of 45,576 individuals were used to identify 8 genetic variants associated with vitamin B 12 levels, serving as genetic instruments for vitamin B 12 status in subsequent analyses. We conducted a Mendelian randomisation (MR)-phenome-wide association study (PheWAS) of vitamin B 12 status with 945 distinct phenotypes in 439,738 individuals from the UK Biobank using these 8 genetic instruments to proxy alterations in vitamin B 12 status. We used external GWAS summary statistics for replication of significant findings. Correction for multiple testing was taken into consideration using a 5% false discovery rate (FDR) threshold. MR analysis identified an association between higher genetically predicted vitamin B 12 status and lower risk of vitamin B deficiency (including all B vitamin deficiencies), serving as a positive control outcome. We further identified associations between higher genetically predicted vitamin B 12 status and a reduced risk of megaloblastic anaemia (OR = 0.35, 95% CI: 0.20-0.50) and pernicious anaemia (0.29, 0.19-0.45), which was supported in replication analyses. Our study highlights that higher genetically predicted vitamin B 12 status is potentially protective of risk of vitamin B 12 deficiency associated with pernicious anaemia diagnosis, and reduces risk of megaloblastic anaemia. The potential use of genetically predicted vitamin B 12 status in disease diagnosis, progression and management remains to be investigated.
Keyphrases
  • genome wide association study
  • machine learning
  • small molecule
  • gene expression
  • dna methylation
  • genome wide
  • deep learning
  • contrast enhanced
  • cross sectional
  • replacement therapy
  • human health