The following is an excerpt of the methods used to generate these data. Full details can be found in the associated publication and supplemental material of Wilder et al. (2020).
Data generation:
Atlantic silverside muscle samples collected from Georgia (GA), New York (NY), Gulf of Maine (GoM) and Gulf of St. Lawrence (GSL) were sequenced by Therkildsen & Palumbi (2017) to a final, average depth of 1.5x across the reference transcriptome. All the samples were collected in the spring. See Therkildsen & Palumbi (2017) and Therkildsen et al. (2019) for further detail. We added 47 additional samples from Oregon Inlet, North Carolina (NC) near Cape Hatteras.
Individually barcoded sample libraries were prepared at Cornells Biotechnology Resource Center using similar methods to Therkildsen & Palumbi (2017), with a few minor modifications. Reagents from the Illumina Nextera kit (96 sample Nextera DNA Library Prep Kit) were used at 1/3 the recommended concentration in 1/10 the recommended volume (5ul instead of 50ul), with 2ng of input DNA. Individual libraries were pooled, and size-selected using a Pippin Prep to remove fragments <286 bp (150 bp insert plus 136 bp of Illumina adapters). Filtered mapped reads for the NC samples had a final average depth of 1.5x.
Sampling and analytical procedures:
Samples were collected directly from the wild as described in Hice et al. (2012).
Locations:
Five locations along the east coast of North America.
USA:Jekyll Island,Georgia 31.02 -81.43
USA:Oregon Inlet,NorthCarolina 35.77 -75.52
USA:Minas Basin,Gulf of Maine 45.2 -64.38
USA:Patchogue,NewYork 40.75 -73.00
Canada:Magdalen Island, Gulf of St. Lawrence 47.4 -61.85
Problem report:
The following samples should be excluded from analysis because of possible errors during library preparation:
JekyllIs_1223, JekyllIs_1224, MagdalenIs_911, MagdalenIs_912, MagdalenIs_913, MagdalenIs_914, MagdalenIs_915, MagdalenIs_916, MagdalenIs_917, MagdalenIs_918, Patchogue_1030