2.3. Study Procedures

The identified banked plasma samples (1000 μL/subject) were retrieved from the genitourinary Gelb tumor bank. These frozen aliquots of plasma were thawed at room temperature and then subjected to high-speed spin. The Qiagen Circulating DNA kit (QIAGEN, Germantown, MD, USA) on the QIAsymphony liquid handling system was used to extract the cfDNA from the plasma samples. ULP-WGS was performed on the extracted cfDNA and sequencing information was run through ichorCNA to detect cases harboring detectable tumor DNA content and CNAs. In detail, the isolated cfDNA was quantified using the PicoGreen (Life Science Technologies, Waltham, MA, USA) assay on a Hamilton STAR-line liquid handling system. CfDNA sequencing libraries were constructed using the Kapa Hyper Prep kit with custom adapters (Integrated DNA Technologies, Coralville, IA, USA). A median of 5 ng of cfDNA input (3–20 ng) was used for ULP-WGS, which was performed using a Hamilton STAR-line liquid handling system (Hamilton Company, Reno, NV, USA). Constructed sequencing libraries were pooled (2 μL of each × 96 per pool) and sequenced using 100 bp paired-end runs over 1× lane on a HiSeq2500 (Illumina, San Diego, CA, USA) for ULP-WGS (~0.1× coverage). The genome was fractioned into T non-overlapping bins of 1 Mb and the HMMcopy Suite1 (http://compbio.bccrc.ca/softwar/hmmcopy/ (accessed on 12 August 2019)) tools were used to count aligned reads based on overlap within each bin. The read counts were then normalized to correct for GC-content and mappability biases using HMMcopy R package. Tumor copy number prediction and tumor DNA estimate were achieved using a hidden Markov Model. The software ichorCNA (available at https://github.com/broadinstitute/ichorCNA (accessed on 19 August 2019)) was used to derive genome-wide copy number plots. Samples passed a quality threshold (median absolute deviation score < 0.115) for accurate purity estimate. Considering the Broad Institute preliminary results [23], up to 40% of samples were estimated to yield >7% tumor purity, which was set as threshold to guarantee the quality of the data. As ichorCNA does not account for subclonal events, to guarantee accuracy, a gene was defined amplified or deleted when ≥5 copies or ≤1 copy, respectively, were found. The analyses were performed by Genomics Platform at the Broad Institute. In order to validate the output achieved with ichorCNA, GISTIC2.0 was rerun on the previously identified sequencing libraries. The GISTIC2.0 module, an evolution of the GISTIC (Genomic Identification of Significant Targets in Cancer) algorithm, identifies probable CNAs by evaluating the frequency and amplitude of observed events [27]. GISTIC was applied to several cancer types [28,29] and aided identifications of several new targets of amplifications and deletions [30,31], and thus was an ideal tool to provide quality metrics for confidence.