Copy number alterations and primary melanoma survival

Our previous work reported a large-scale copy number (CN) study of primary melanoma [1]. Next generation sequencing (NGS) data from 303 formalin fixed paraffin embedded (FFPE) samples from the Leeds Melanoma Cohort (LMC) were generated. Libraries were generated by random shearing and then sequenced (1.7x coverage). In this study, problematic regions and common germline variations in the genome were identified and excluded accounting to approximately 13.5 % of the autosomal genome. CN was generated by read count accumulated into 10k bp windows, adjusted jointly for sequence mappability and GC-content and in comparison, with Caucasian genomes (n=312) from the 10k Genome Project [2, 3]. GISTIC 2.0.23 identified significantly deleted or amplified regions in the genome [4]. Comparisons with the TCGA data showed high similarity between the two datasets in terms of location and proportion of samples with CN changes [5, 6]. Minor difference in terms of frequency which may be due to the type of samples processed (frozen vs FFPE), platform used (NGS vs SNPS Array), or disease stage (primary vs metastatic) offer opportunity for discovery of novel CNA in melanoma
Three measures of overall genome instability namely Fraction of Genome Altered (FGA), Aneuploidy Score (AS), and a devised metric I referred as Mean Weighted Segment Mean (MWSM) were estimated. MWSM showed strongest association to most of the patient clinical characteristics and survival among the three measures. Focal analysis was done using each 10k window level copy number data which allowed detection of small copy number aberrations that are associated with patient clinical characteristics and survival
In conclusion, this study showed the feasibility of extracting and analysing whole genome copy number data from FFPE samples given that satisfactory amount of quality control steps to improve data quality is done. While there were interesting associations identified between copy number and patient clinical and tumour characteristics, validation of these results once similar population-based study cohort becomes available.