For validation purposes, five liver extracts with low recoveries were diluted up to 1000 times and analyzed on a Xevo TQ-S mass spectrometer (Waters Corporation, Milford, USA), which is a more sensitive instrument compared to the Quattro Premier find more XE. The recoveries of 13C4-PFOS increased from 10–44% to 36–80% in the × 100 and × 1000 diluted samples (Fig. S1, Supplementary data). To compare PFOS concentrations in undiluted (u) and diluted (d) extracts, the mean normalized difference (%) was calculated using the formula: ((u − d) / ((u + d) / 2) × 100). The calculated concentrations

of all the diluted extracts, except for one sample, were well in range with the initial concentrations (average mean normalized difference of 18%). Consequently, reliable results can be produced even when recovery rate is low since the internal standard and the native compound are equally suppressed. Recoveries, method reproducibility and method detection limits (MDL) for all samples are presented

in Table S1, Supplementary data. One milliliter of ultra pure water was used as procedural blanks and extracted in the same way as the real samples. The MDL was defined as the mean concentration in the procedural blanks plus three standard deviations, and the limit of detection (LOD) for individual samples was calculated as three times the noise level. Overall good recoveries (> 50%) of 13C-PFOS and 13C-PFOA were measured for the samples after the replacement of the recovery standard 7H-PFHPA

to 13C8-PFOS and 13C8-PFOA in the middle of the project. One two year old mink TSA HDAC mouse caught in autumn in the G area was excluded due to non-reproducible results of the diluted extracts. Using the general linear model (GLM) procedure of SAS (SAS Institute Inc., Cary, NC, USA, version 9.02.01), a multiple regression model with the concentrations Temsirolimus mouse of PFHxS, PFOS, PFNA, PFDA or PFUnDA as dependent variable and the sample area, sample season, age, body condition, year (of capture) and body weight as independent variables were elaborated on. PFBS, PFOA, PFDoDA and PFTrDA were excluded from this model since the concentrations were consistently low (< 17 ng/g). The model was fitted manually, starting with all variables in the model. Variables that were unsignificant (p > 0.05) for all dependent variables were removed. Relevant interactions between the effects were tested but none were included in the model due to insignificance or small sample size. The variable age was tested in several ways (different assignments into categories and numerical approaches), but had no significant effect. Area and season were the only variables that had a significant effect and were therefore the only variables kept in the final model: Y=μ+AREA+SEASON+ERROR.Y=μ+AREA+SEASON+ERROR.