Our analysis does not support the previously described association between the 10q risk locus and FGFR2 levels (P = 0.537). Prior studies have yielded inconsistent results with 3 studies demonstrating an association and 2 that have not (Huijts et al., 2011; Martin et al., 2011; Meyer et al., 2008; Riaz et al., 2012; Sun et al., 2010). Variability between the studies may be due to different sample sizes, different tissue types (normal, tumor), and/or measurement of different isoforms. The largest study was performed on 1,401 breast tumor samples and the results showed no association. eQTL studies performed in different cell types and across the full spectrum of expressed isoforms will help to further clarify these observations.
