Gadus morhua

All read datasets originated from DNA extracted from the same individual fish, designated NEAC_001, a wild-caught male specimen of the North-East Arctic population, sampled with the main purpose for sequencing initiative of the Atlantic cod genome and described in detail in [5 ]. We always strive to limit the effect of our sampling needs on populations and individuals. This individual was sampled in connection with a research survey conducted by Norwegian Institute for Water Research as part of part of larger hauls for stock assessments. The fish were humanely sacrificed by administration of other sedatives before sampling in accordance with the guidelines set by the ’Norwegian consensus platform for replacement, reduction and refinement of animal experiments’ (www.norecopa.no). See Additional file 1: Table S1 for an overview of different DNA datasets generated from this individual.
Roche/454 reads were sequenced as described previously [5 ]. The Roche/454 software gsRunProcessor version 2.6 was used to redo basecalling for all sequencing runs generated for the NEAC_001 sample [5 ].
One hundred eighty bp insert size and 300 bp insert size libraries were constructed with Illumina DNA paired end sample preparation reagents and sequenced at the Norwegian Sequencing Centre. The 5 kbp insert size libraries were prepared with the Illumina Mate Pair gDNA reagents and sequenced at the McGill University and Génome Québec Innovation Centre. All Illumina libraries were sequenced on the HiSeq 2000 using V3 chemistry 100 bp paired end reagents.
PacBio SMRT sequencing was performed on PacBio RS instrument (Pacific Biosciences of California Inc., Menlo Park, CA, USA) at the Norwegian Sequencing Centre (www.sequencing.uio.no/) and at Menlo Park. Long insert SMRTbell template libraries were prepared at NSC (10 kbp insert size) and Menlo Park (22 kbp insert size) according to PacBio protocols. In total, 147 SMRT-cells were sequenced using C2 and XL polymerase binding and C2 and XL sequencing kits with 120 min acquisition. Approximately 7.6 Gb of library bases were produced from 10 kb SMRTbell libraries sequenced on 102 SMRT cells using C2/C2 chemistry (average polymerase read length of 3 kb). The 22 kb SMRTbell library was sequenced using C2/XL (22 SMRT cells, average polymerase read length of 4.5 kb) and XL/XL (23 SMRT cells, average polymerase read length of 5 kb) chemistry producing 5.5 Gb of library bases.

To detect variations in the L. crocea genome, we chose nine species (Larimichthys crocea, Gasterosteus aculeatus, Takifugu rubripes, Tetraodon nigroviridis, Oryzias latipes, Gadus morhua, Danio rerio, Gallus gallus, and Homo sapiens). Proteins that were greater than 50 amino acids in size were aligned by BLAST (-p blastp-e 1e-7), and Treefam [70 (link)] was used to construct gene families for comparison.
The 2,257 single-copy genes from the gene family analysis were aligned using MUSCLE [71 (link)], and alignments were concatenated as a single data set. To reduce the error topology of phylogeny by alignment inaccuracies, we used Gblock [72 (link)] (codon model) to remove unreliably aligned sites and gaps in the alignments. The phylogenetic tree and divergence time were calculated using the PAML 3.0 package [19 (link)].
Gene family expansion and contraction analyses were performed by cafe [73 (link)]. For optical, olfactory receptor, and auditory system-related genes, we downloaded the genes from Swissprot or Genebank and predicted their candidates using BLAST and Genewise to determinate copy numbers. Pseudogenes produced by frame shift were removed. Phylogenetic analysis of the expanded gene families was based on maximum likelihood methods by PAML 3.0 [19 (link)], and the phylogenetic tree was represented by EvolView [74 (link)].
Amino acid sequences from six representative teleosts (Larimichthys crocea, Gasterosteus aculeatus, Danio rerio, Oryzias latipes, Takifugu rubripes, and Tetraodon nigroviridis) were aligned by BLAST (-p blastp-e 1e-5-m 8), and reciprocal-best-BLAST-hit methods were used to define orthologous genes in six teleost fish. Because alignment errors are an important concern in molecular data analysis, we made alignments of codon sequences, which are nucleotide sequences that code for proteins, using the PRANK aligner [75 (link)]. Positive selection was inferred, based on the branch-site K_a/K_s test by codeml in the PAML 3.0 package [19 (link)].

Approximately 10 g samples of muscle (±0.0001 g) in duplicate were dried to constant weight at 105 °C. Then, the samples were combusted at 480 °C for 12 h using laboratory furnaces Nabertherm P330) (Nabertherm GmbH, Lilienthal, Germany). The white ash was dissolved in 1 M HNO₃ (Suprapur-Merck, Darmstadt, Germany). Each sample was then quantitatively transferred using deionized water (Merck-Millipore Elix Advantage 3, Burlington, MA, USA) into a volumetric flask of volume 25 mL.
The contents of these metals, except for cadmium, sodium, and potassium, were determined using atomic absorption spectrometry (iCE 3500 Series AAS, Thermo Scientific, Waltham, MA, USA), using cathode lamps appropriate for the given elements and background correction (a deuterium lamp). The absorption technique (acetylene-air flame) was applied to determine the content of Zn, Cu, Fe, Mn, Ca, and Mg. When determining calcium to eliminate the influence of phosphorus, the solution of lanthanum chloride was added (0.5%La⁺³) to all samples and standards. The absorption wavelengths were as follows: 248.3 nm for iron, 213.9 nm for zinc, 324.8 nm for copper, 279.5 nm for manganese, 285.2 nm for magnesium, and 422.7 nm for calcium. The detection limits (LOD) were 0.125 mg/kg for Fe, 0.1 mg/kg for Zn, 0.05 mg/kg for Cu, 0.05 mg/kg for Mn, 0.025 mg/kg for Mg, and 0.5 mg/kg for Ca.
The flame emission technique (acetylene-air flame) was used to determine sodium and potassium content at 589.0 nm and 766.5 nm, respectively. Atomic absorption spectrometry (iCE 3500 Series AAS, Thermo Scientific, Waltham, MA, USA) was used for this purpose. The detection limits (LOD) were 1 mg/kg for Na and 5 mg/kg for K.
The concentration of cadmium was measured using flameless atomic absorption spectrometry FAAS (Thermo Scientific iCE 3500, Waltham, MA, USA) with ZEEMAN background correction and atomization in a graphite cuvette. The cadmium determination was carried out under the following conditions: absorption wavelength for cadmium—228.8 nm; lamp current—50%; slit—0.5 nm; sample volume—20 μL; modifier—Mg(NO₃)₂; dry—100 °C; ash—600 °C; and atomize—1000 °C. The detection limit (LOD) was 0.00007 mg/kg.
Four blanks and four standards were analyzed with each batch of samples. Calibration curves were prepared using four standard solutions (1000 μg/L) with 0.1 M HNO₃ (J.T.Baker^®, Netherlands). The calibration curves were linear within the range of metal concentration (regression coefficients R² ≥ 0.999).
The quality control of methods was tested using the reference material BCR CRM 422 (lyophilized muscle tissue of cod, Gadus morhua (L.)) with a certified value of mercury. The recovery rates were 105.0% Zn, 103.0% Cu, 96% Fe, 103% Mn, 102.9% Cd, and 100.2% Hg, respectively (n = 4).

A total of 45 dried salted cods (2–3 kg of dry weight; n = 15 samples/origin) were used in this study. The Atlantic cod was fished in the Atlantic northeast (FAO 27 area) within the Exclusive Economic Zones (EEZ) of Norway (n = 15) and Iceland (n = 15), while the Pacific cod was caught in the Pacific northeast (FAO 67 area) within the Alaska EEZ (n = 15). The dried salted cods used herein were randomly selected among specimens harvested between January and April of 2018, sharing analogous fishing and processing methodologies, and were all collected at Riberalves Company (Carvalhal, Torres Vedras, Portugal).
According to Good Manufacturing Practices of Riberalves, the average time between cod harvesting and the salting and drying processes was two to three months. Salting and drying processes were performed according to commercial procedures for specimens weighing 2–3 kg.
All samples for the various analyses were taken from the central portion of loins after the five to six days of the drying process had been completed. Skin and bones were manually removed, and the remaining part of muscle tissue (designated hereafter as the edible portion) was cut into thin slices before blending in a food processor (Moulinex, France). Half of the homogenised material was vacuum packed and frozen/stored at −20 °C until further analysis, within one month, while the other half was freeze dried (−60 °C and 2.0 h Pa) up to constant weight, using an Edwards Modulyo freeze dryer (Edwards High Vacuum International, West Sussex, UK), packed and kept at room temperature, and analysed within one month. In all assays, the 45 samples were analysed in duplicate, and the results were accepted when the coefficient of variation between duplicates was below 5%. Whenever duplicates were associated with a higher coefficient of variation, the analytical procedure was repeated. The average value obtained for each sample analysis was used in the statistical analysis (n = 45; 15 samples/cod group).