Library Preparation for Sequencing
We identified three key parameters in the Therkildsen and Palumbi (2017) method that we could optimize in order to target the ideal fragment distribution and avoid loss of critical DNA when working with feather DNA samples. The three key parameters modified herein were: (1) the ratio of tagmentation transposome (which cleaves DNA and adds an adapter for indices) to input DNA quantity, (2) the duration of tagmentation incubation, and (3) the duration of the indexing PCR elongation time. We identified these parameters as key because: (1) the ratio of tagmentation transposome to DNA molecules determines the average fragment length, (2) the duration of tagmentation incubation determines the efficacy of fragmentation by allowing enough time for tagmentation to occur, (3) the indexing PCR elongation time dictates the average fragment length of amplified DNA. We conducted initial tests to determine the most efficient modifications to the protocol (not presented in this manuscript) and finalized the protocol using those adjusted parameters.
In order to assess the efficiency of our method, we extracted DNA from 50 flight feather calamus tips of the American Redstart stored at -20 oC for variable lengths of time. DNA was extracted immediately prior to library preparation. For extractions from feathers, like other low-yield samples, maximizing gDNA yield is critical. Therefore, we followed the commercially available DNEasy Blood and Tissue Extraction Kit (Qiagen) protocol but with the following modifications. To each sample, we added 10 µL of 1 M Dithiothreitol (DTT) to the initial lysis step to aid in breaking down disulfide bonds found in the keratin of feathers. Flowthrough after the first filtration step when lysate was transferred to the spin column was pipetted back onto the filter for a second centrifugation. Prior to the final elution step, AE buffer was placed in an incubator at 56 oC. During the final elution step, AE buffer was left to incubate on the filter for five minutes instead of two. We eluted feather extractions into 400 µL (two rounds of 200 µL elutions through the spin column as recommended by Qiagen protocol for maximum yield). Prior to proceeding with library prep, we concentrated feather DNA extractions using a 1:1 ratio of Serapure beads (Faircloth and Glenn 2014) from 400 µL to 15 µL and eluted into 10mM Tris-Hcl (Fig. 1, step 1a).
To ensure that final library concentrations will be similar across samples, we quantified each DNA extraction using a Qubit dsDNA High Sensitivity Assay Kit (Invitrogen) and normalized each sample to a concentration of 0.48 ng/µL- 4.5 ng/µL, with a target of 2.5 ng/µL (Fig. 1, step 1b). To fragment the DNA and “tag” it with Nextera adapters, we added 2.50 µL of TD Buffer and 0.5 µL of TDE1 Enzyme (Illumina) to 1 µL of normalized DNA (e.g. 0.48 ng to 4.5 ng of DNA) and incubated the samples in a thermocycler at 55 oC for 20 minutes (Fig. 1, step 2).
To amplify the tagmented DNA and add Nextera indexing adapters for sequencing, we pipetted 1 µL of each index primer into the appropriate well of tagmented DNA until all samples had a unique pair of dual indexes. We then added 6.0 µL of Kapa Hifi Hotstart Mix (KMM; Kapa Biosystems) before running in a thermocycler as follows: held at 72o for 3 minutes, held at 98o for 2 minutes and 45 seconds, cycled 8 times through 98o for 15 seconds, then 62o for 30 seconds, then 72o for 30 seconds, held at 72o for 1 minute, and then held at 4o until removed from thermocycler (Fig. 1, step 3). As per the Therkildsen and Palmubi (2017) method, this indexing PCR had more cycles than the original Illumina protocol and was broken into two stages (Indexing PCR and Reconditioning PCR). Additional cycles were added in the Therkildsen and Palumbi (2017) method because the tagmented DNA was not purified prior to the Indexing PCR, making the PCR reaction less efficient. To further amplify copies of indexed DNA without using additional Nextera indices, we added 7.6 µL of KMM, 4.4 µL of ultrapure water, and 1.6 µL each of a custom 10uM primer pair (P1 = AATGATACGGCGACCACCGA; P2 = CAAGCAGAAGACGGCATACGA) to each library. We ran the samples in a thermocycler as follows: held 95o for 5 minutes, cycle 4 times through 98o for 20 seconds, then 62o for 20 seconds, then 72o for 2 minutes, hold at 72o for 2 minutes, and then held at 4o until removed from the thermocycler (Fig. 1, step 4).
To purify the PCR product and remove undesirable fragments, we followed standard Ampure bead protocol (Beckman Coulter) using a 0.7:1 bead to DNA ratio which will remove below approximately 320 bp and eluted into 30 µL of 10mM Tris-Hcl (Fig. 1, step 5). In order to avoid overrepresentation of one individual during whole genome resequencing, we then quantified using a Qubit dsDNA High Sensitivity Assay Kit (Invitrogen) and pooled an equal number of copies of each sample into a 1.5 mL tube (Fig. 1, step 6). Finally, in order to increase the final concentration of the pooled libraries and increase sequencing efficiency, we then followed the standard Ampure double-sided size selection protocol, using a 0.63:1 bead to DNA ratio to remove large fragments and a 0.73:1 bead to DNA ratio to remove small fragments, and eluted into 30 µL of 10mM Tris-HCl (Fig. 1, step 7). After the pooled library has been concentrated and double size selected, we perform final quality control (QC) with Qubit quantification and Tapestation 2200 fragment distribution analysis (Agilent) (Fig. 1, step 8).
Using the above method (for full written protocol, see Online Resource 1), we prepared one WGS library from American Redstart (Setophaga ruticilla) samples for low coverage WGS (< 2x). The library was prepared with 50 unique feather samples, extracted from a single flight feather calamus, with starting DNA concentrations of 0.48–4.5 ng/µL. The library was sequenced on one full 2 x 150 bp PE (paired end) HiSeq 4000 lane (Illumina).
Bioinformatic Analysis and Quality Checking
We trimmed sequence adapters using the program TrimGalore version 0.6.5 (https://github.com/FelixKrueger/TrimGalore), a wrapper for Cutadapt (Martin, 2011). We used the Burrows-Wheeler Aligner software version 0.7.17 (Li & Durbin, 2009) to map reads to an assembled genome of the yellow warbler (Setophaga petechia; Bay et al. 2018). After mapping, the resulting SAM files were sorted, converted to BAM files, and then indexed using Samtools version 1.9 (Li et al., 2009). We marked and removed read duplicates with MarkDuplicatesSpark from GATK version 4.1.4.0 (McKenna et al., 2010). Depth of sequencing coverage was calculated using the depth function in samtools (Li et al. 2009). We checked the sequence quality of the BAM files using the flagstat function from samtools to determine the proportion of reads that were properly mapped and paired.