Supporting dataset on the validation and verification of the analytical method for the biomonitoring of 360 toxicologically relevant pollutants in whole blood

The dataset presented in this article supports “Micro QuEChERS-based method for the simultaneous biomonitoring in whole blood of 360 toxicologically relevant pollutants for wildlife” (Rial-Berriel et al., 2020). The supplementary data are: (1) Detailed validation data of the LC-MS/MS and GC-MS/MS methods for the quantification of 360 chemicals covering bias and precision (intra- and inter-day variability) for retention times, linearity, and limits of quantification. (2) Graphical data of the matrix effects on the quantification of all of the analytes. (3) Individual data of the 51 chemicals detected in real whole blood samples from two raptor species: 36 barn owls (Tyto alba) and 112 common kestrels (Falco tinnunculus).

© 2020 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY license.

Specific subject area
Analytical chemistry applied to biological samples to perform biomonitoring of environmental pollutants Type of data Tables and figures How data were acquired Ultra-high performance liquid chromatography tandem coupled to triple quadrupole mass spectrometry (LC-MS/MS). Agilent Technologies (Palo Alto, USA) models 1290 (UHPLC) and 6460 (MS/MS). Gas chromatography tandem coupled to triple quadrupole mass spectrometry (GC-MS/MS). Agilent Technologies (Palo Alto, USA) models 7890B (GC) and 7010 (MS/MS).

Data format
Analyzed: • Extracted and analyzed LC-MS/MS and GC-MS/MS data for the validation studies • Quantified chromatogram data of pollutants in the blood of barn owls and common kestrels Raw data: • All raw data corresponding to the 5 replicates of each concentration tested in the validation experiments presented in Table 1.
• All individual quantitative data obtained after application of the method on 148 common kestrels and 39 barn owls presented in Tables 2-5.
• All raw data used for the elaboration of Fig. 1.

Parameters for data collection
The validation data of the developed method included, limit of quantification (as the lowest calibrator that fulfilled de validation criteria), linearity, accuracy (expressed as % bias and precision (intra-and inter-day RSDs)) for each of the 360 analytes. The matrix effect on the quantification of the analytes is graphically represented as percentages. For clarity the results are expressed as relative percentage. The quantification data were obtained analyzing a series of 148 blood samples, obtained from a field ecology work in nest boxes of barn owls ( Tyto alba , n = 36), and common kestrels ( Falco tinnunculus , n = 112).

Description of data collection
MassHunter Quantitative Analysis was employed to collect and analyze the chromatographic data delivered by the triple quadrupole mass spectrometers coupled to both UHPLC and GC. The linearity was assessed by injecting a 12-point calibration curve prepared in the blank matrix and extracted with the developed micro-QuEChERS method. To test bias and precision, standard solutions of 360 reference standards were employed to spike blank whole blood samples at five concentration levels (0.1, 0.5, 1, 5, and 20 ng/ml) were injected in quintuplicate. The bias and repeatability (intra-day variability) were determined by those quintuplicate analyses of each sample, as these were injected within 24 h. The reproducibility (inter-day variability) was measured on three non-consecutive days within a two-week span. The matrix effect was assessed by extracting enough amount of blank matrix with the developed method and fortifying these extracts with three levels of the mixture of 360 chemicals (0.2, 2, and 10 ng/ml), and quantified against a calibration curve prepared in the solvent (1% FA-acetonitrile). Regarding the quantitative data, the real samples were prepared with the developed methodology and analyzed by UHPLC and GC. Value of the data • An easy way to consult the validation data of the methodological development detailed in the main article is provided. • Additionally, the validation data might serve as a reference to other researchers developing methods in complex biological matrices. • The detail of the matrix effect on each of the 360 compounds analyzed may allow other researchers to decide whether they need to prepare their calibration lines in matrix or in solvent. • Data on numerous environmental pollutants are presented for the first time for two apex predators: barn owl ( Tyto alba ) and kestrel ( Falco tinnunculus ). • The biomonitoring data presented can be used by other researchers in ecotoxicology for comparison purposes, and for meta-analyses of chemical contamination in birds of prey. Table 1 shows the major validation parameters of each of the 360 chemicals optimized in this methodology. Validation criteria fulfilled those established in the SANTE guidelines [1] . Given the particularities of the matrix -whole blood -the guidelines of the SWGTOX were also considered, especially concerning the preparation of the matrix-matched calibration curve [2] . The data reported in this table complement those included in Table 1 of the article by Rial-Berriel et al. [3] , where the parameters of identity and selectivity of each compound are shown. Now, in this table, the data of linearity (as a working range), the LOQ (set as the lowest point of the calibration curve that complied identity, bias and precision criteria), repeatability (as intraday RSD), and reproducibility (as interday RSD) are shown for each of the chemicals. The bias and precision (repeatability and reproducibility) data are presented for five fortification levels (0.1, 0.5, 1, 5, and 20 ng/ml). Raw data of the five replicates for each concentration are provided in the supplementary file 1. Fig. 1 shows the effect that the components of the matrix have on the quantification of each chemical substance, to demonstrate the need to perform said quantification with calibrators prepared in a white matrix for at least 45% of the compounds. Results are shown as a relative percentage quantified by ACN calibration curve. When the differential is greater 20% or −20%, matrix interference is considered to exist. For clarity, the compounds are identified numerically, from 1 to 360, as they are numbered in Table 1 . Raw data of the six charts of this figure are provided in the supplementary file 2. Tables 2 and 3 show the quantitative results of the contaminants found in 36 barn owls ( Tyto Alba ) and Tables 4 and 5 , show those found in 112 common kestrels ( Falco tinnunculus ). The results of persistent and non-persistent organic pollutants are presented separately. For clarity, only contaminants that have tested positive for at least one individual are shown. Raw data of the quantification of the 360 pollutants in these 187 birds are provided in the supplementary file 3.

Chemicals, reagents, and calibrators
Certified pure standards of the 360 chemicals included in the methodology (Dr Ehrestorfer, Augsburg, Germany; CPA Chem, Stara Zagora, Bulgaria; A2S -Analytical Standard Solutions, Staint Jean D'Illac, France; Sigma-Aldrich, Augsburg, Germany; Accustandard, New Haven, USA; ( continued on next page )   ( continued on next page )      This working solution was employed to prepare 12-point calibration curve, either in 1% FA-ACN or in whole blood (a 1:1 mixture of chicken and goat blood, obtained from healthy animals in the Veterinary Faculty of the University of Las Palmas de Gran Canaria). The intermediate solution was also employed for preparing quality controls (QC) and all the fortification levels in blank matrix and extracted blank matrix, that were employed in the validation experiments. The salts for the QuEChERS extraction according to the AOAC method were acquired in commercial premixes from Agilent Technologies (Palo Alto, USA). Acetonitrile (ACN), methanol (MeOH), and FA were of the maximum purity available and were purchased from Honeywell (Charlotte, USA). Ammonium acetate was from Fisher (Fisher Scientific UK, Loughborough, UK). The water was prepared in the laboratory using an ultra purification system (Millipore, Molsheim, France). Table 3 Concentrations of non-persistent organic pollutants found in whole blood (ng/ml) of a series of barn owls sampled in Castilla-Leon (Spain) during 2018 and 2019.

Sample preparation
Two hundred fifty microliters whole blood from wild birds or blank matrix (goat + chicken whole blood) either fortified with the 360 analytes plus the internal standards (ISs), the ISs alone, or without fortification, were subjected to an AOAC-QuEChERS extraction method (onestep, without cleanup). The mixture of IS (acenaphthene-d10, chlorpyrifos-d10, chrysene-d12, diazinon-d10, PCB 200, and phenanthrene-d10 for the GC method, atrazine-d5, carbendazim-d3, cyromazine-d4, diazinon-d10, linuron-d3, and pirimicarb-d6 for the LC method) was added to all the tubes before the extraction. The samples were orbital shaken for 1 h, and then, 500 μl of 1% FA-ACN were added, before tubes were placed in an ultrasonic bath at room temperature for 20 min. Then, the QuEChERS salts (150 mg anhydrous magnesium sulfate and 37.5 mg sodium acetate) were added and, the tubes were vortexed 30 s and vigorously manually shaken Table 4 Concentrations of persistent organic pollutants found in whole blood (ng/ml) of a series of common kestrels sampled in Castilla-Leon (Spain) during 2018 and 2019.    for 60 s. After a 5 min centrifugation, the supernatant was filtered (0.2 μm) and used directly for chromatographic analyses. The optimization of the procedure is detailed in the main article [3] .

Chromatographic analyses
Two complementary analyses were needed for the quantification of 360 environmental pollutants in whole blood. Thus, 234 compounds were analyzed by liquid chromatography and other 126 chemicals were analyzed by gas chromatography. Both techniques were tandem coupled to triple quadrupole mass spectrometry (LC-MS/MS and GC-MS/MS). All the equipment employed was from Agilent Technologies (Palo Alto, USA): UHPLC 1290 coupled to 6460 mass spectrometer for LC, and a 7890B GC coupled to 7010 mass spectrometer for GC. The chromatographic separations in LC were performed using an InfinityLab Poroshell 120 (2.1 mm × 100 mm, 2.7 μm). The stationary phase in GC consisted of two fused silica ultra-inert capillary columns Agilent HP-5MS (15 m × 0.25 mm i.d., 0.25 μm film thickness), that were connected by a purged union, to allow the backflushing. All the conditions in which these apparatus were operated, as well as the optimization procedure, are described in detail in the main article [3] .

Validation
This article provides supporting information on the data of the validation process, which was carried out according to the criteria established in the SANTE and SWGTOX guide [ 1 , 2 ] and, taking into account our previous experience in developing and validating chromatographic methods in complex biological matrices [4][5][6] . Although initially the experiments were carried out with chicken and goat blood separately, we found no significant differences. Therefore, we decided to use a 1: 1 mixture of both blood types to complete all validation experiments, following the recommendations of the SWGTOX guide. All the parameters of the validation process are shown in this article, except for the identity and selectivity parameters, which are shown in Table 1 of the main article [3] , and carryover, which is also explained in the main article. As the blank matrix was not completely free of all of the contaminants, the signal of the blanks was subtracted from fortified samples in all the experiments.
Linearity was assessed within the range of concentrations that were considered appropriate for the purpose of biomonitoring (0.1 to 20 ng/ml). All the calibrators (12 points) were individually prepared in blank matrix in quintuplicate by adding an appropriate volume of fortification solution. Only compounds that showed a correlation coefficient ( R 2 ) higher than 0.93 were kept in the method.
The accuracy (bias and precision) was calculated for all the 12 fortification levels injected in the chromatographic systems in quintuplicate. The bias is expressed as the percentage of the theoretical level of fortification. As recommended in the guidelines, only compounds with recoveries between 70% and 120% were kept in the method. However, due to their importance for biomonitoring certain exceptions were admitted, including some compounds with recovery percentages below or above these limits, but which were highly reproducible. Precision (repeatability and reproducibility) is expressed in terms of relative standard deviation (RSD) of the different replicates. For the reproducibility of the method, only 5 levels of fortification (0.1, 0.5, 1, 5, and 20 ng/ml) were prepared in triplicate on three different days, within a period of 2 weeks. Therefore, the RSDs were calculated from 9 values. Only those compounds in which the RSD values were less than 20% remained in the method. In Table 1 , we show the data of these 3 parameters only for the five levels of fortification mentioned above, although the rest also met the specified criteria.
The LOQ of this methodology was calculated over five runs of fortified blank matrix samples of three different sources (chicken, goat, and a mixture of both), as recommended (Scientific Table 5 Concentrations of non-persistent organic pollutants in whole blood (ng/ml) of a series of common kestrels sampled in Castilla-Leon (Spain) during 2018 and 2019.   Working Group for Forensic, 2013). The lowest non-zero calibrator approximation was employed to calculate de LOQs. This means that the lowest point of the calibration curve that complied identity, bias and precision criteria was set as the LOQ for a given compound. All compounds with LOQ > 5 ng/ml were eliminated from the method since these levels are not considered adequate for biomonitoring studies. The calculated LOQs are shown in Table 1 .
The influence of the matrix components on the performance of the method was evaluated by applying the extraction method to a sufficient quantity of blank whole blood to produce a blank matrix extract, which was subsequently fortified at three levels for the mixture of 360 chemicals (0.2, 2, and 20 ng/ml), and quantified against a calibration curve prepared in the solvent (1% FAacetonitrile). The data corresponding to these experiments are shown in Fig. 1 .