On 2017 Feb 09, Karsten Suhre commented:

Very interesting paper - The TXNIP cg19693031 association with Type 2 Diabetes clearly replicates everywhere. The authors write "While finalizing this manuscript,three studies were published and observed an association of DNA methylation at cg19693031 in TXNIP, and type 2 diabetes (36–38)." Actually, there is a fourth paper that came out in Jan 2016: "Epigenetic associations of type 2 diabetes and BMI in an Arab population" by Al Muftah et al. (Clin Epigenetics. 2016 Jan 28;8:13. doi: 10.1186/s13148-016-0177-6, PMID: 26823690). You may be interested to take a look at Table 4 of the Al Muftah paper - it puts the recent Petersen et al. metabolomics findings (ref 35 in the present Soriano-Tárraga et al. paper) into the context of the TXNIP association: The associated metabolic phenotype (metabotype) of the TXNIP CpG association is characteristic of a diabetes state. What is more, the obesity (BMI) associated CpG loci cg06500161 (ABCG1) and cg00574958 (CPT1A) also display the same diabetes metabotype. This suggests that the methylation of all of these sites could be a regulatory response to obesity induced deregulated metabolism. It would be interesting to test whether the TXNIP CpG association holds in a cohort of non-obese diabetics.

On 2017 Feb 09, Karsten Suhre commented:

Very interesting paper - The TXNIP cg19693031 association with Type 2 Diabetes clearly replicates everywhere. The authors write "While finalizing this manuscript,three studies were published and observed an association of DNA methylation at cg19693031 in TXNIP, and type 2 diabetes (36–38)." Actually, there is a fourth paper that came out in Jan 2016: "Epigenetic associations of type 2 diabetes and BMI in an Arab population" by Al Muftah et al. (Clin Epigenetics. 2016 Jan 28;8:13. doi: 10.1186/s13148-016-0177-6, PMID: 26823690). You may be interested to take a look at Table 4 of the Al Muftah paper - it puts the recent Petersen et al. metabolomics findings (ref 35 in the present Soriano-Tárraga et al. paper) into the context of the TXNIP association: The associated metabolic phenotype (metabotype) of the TXNIP CpG association is characteristic of a diabetes state. What is more, the obesity (BMI) associated CpG loci cg06500161 (ABCG1) and cg00574958 (CPT1A) also display the same diabetes metabotype. This suggests that the methylation of all of these sites could be a regulatory response to obesity induced deregulated metabolism. It would be interesting to test whether the TXNIP CpG association holds in a cohort of non-obese diabetics.