Bariatric surgery is currently the most effective treatment for type 2 diabetes (T2DM), in some cases inducing complete remission within days of surgery. The mechanisms driving this phenomenon remain unclear. This thesis describes the collection of clinical and genomic information from participants of the Personalised Medicine for Morbid Obesity (PMMO) study undergoing Roux-en-Y gastric bypass or vertical gastrectomy, investigates the ability of clinical and genomic factors to predict diabetes remission following surgery, and identifies early post-surgery changes in gene expression that may drive remission.

Although most T2DM improves to some extent following bariatric surgery, not everyone achieves remission. Scoring systems to predict remission help to ensure that limited NHS resources are directed towards those most likely to benefit: an early aim was to explore how these might be improved. Two such scores, DiaRem and Advanced-DiaRem, were applied to the PMMO cohort, along with the individual clinical factors of T2DM duration and number of prescribed T2DM drugs. Advanced-DiaRem had greatest predictive value in this heterogeneous cohort, particularly in non-insulin-treated participants for whom remission prediction scores have generally decreased efficacy. Identification of genomic predictors of diabetes remission may reveal more about the aetiology of diabetes and why people respond differently to bariatric surgery. Using RNA-sequencing and quantitative polymerase chain reaction (qPCR) a transcript of HMGCS1 (required for de novo cholesterol biosynthesis) was identified as a genomic predictor of T2DM remission after Roux-en-Y gastric bypass. This is the first transcriptomics biomarker identified as predictive of T2DM remission from pre-surgery peripheral blood.

I also describe the first study comparing early post-operative transcriptomic profiles between remission and non-remission groups. Transcripts of ANXA6 (which regulates intracellular cholesterol distribution, further implicating bile-acids in the mechanisms of remission) and TRAF1 (which regulates NF-kB pathways) were differentially expressed in RNA-sequencing data, but this was not confirmed by qPCR analyses.