In 2016, Baltimore lost 454 people to heroin-related overdoses, up from 260 the previous year, and 419 people to fentanyl-related overdoses, up from 120 in 2015 [56, 57]. The figures available for Massachusetts do not distinguish between heroin and prescription opioids. In 2016, among the 1374 individuals whose deaths were opioid-related (including heroin) and a toxicology screen was also available, 1031 of them (75%) had a positive screen result for fentanyl, an increase from 754 (57%) in 2015, although this may depend on the frequency of toxicological screening [58, 59]. Drug overdose deaths in New Hampshire increased by 1629% between 2010 and 2015, largely as a result of fentanyl. Hillsborough County, the location of Nashua, where 43.6% of the fentanyl deaths occurred, was most affected by these overdose deaths. The rate of death caused by fentanyl, heroin, and other opioids rose sharply between 2015 and 2016 [60]. In Chicago in 2016, there were 487 overdose deaths involving heroin and 420 involving fentanyl, both rising from the previous year [61]. In San Francisco in 2016, 41 deaths were attributed to heroin overdose and 22 attributed to fentanyl, doubling from the previous year [62, 63]. Data for 2017 are not available for all sites, but Baltimore showed a small decline in heroin-related deaths (from 334 in January to September 2016 to 305 in the same period of 2017) but a much larger increase in fentanyl-related deaths (from 276 January to September 2016 to 427 in the same period in 2017) [64]. Massachusetts experienced a modest decline in overall opioid deaths in 2017 but an increase in the proportion screening positive for fentanyl (to 83%) [65].
In Chicago, where the powder heroin can be snorted rather than smoked, several injectors described snorting their heroin before injecting. Ray, in his 50s and using for 25 years, explained that this not only gave an indication of its strength but also a taste at the back of the throat which, he believed, was indicative of its ingredients:
A number of genomic properties (such as replication timing, transcriptional activity and chromatin state) influence the density of point mutations30,31 and copy-number alterations32, but how this relates to individual classes of structural variant is unclear. From the literature, we compiled a library of the genome-wide distribution of 38 features including replication timing, GC content, repeat density, gene density and distance to G-quadruplex motifs, among others. Replication timing had the strongest association with the occurrence of structural variants; deletions are enriched in late-replicating regions, and tandem duplications and unbalanced translocations occur preferentially in early-replicating regions (Fig. 5c, Extended Data Fig. 8). For individual patients with high numbers of deletions or tandem duplications, we observed notable heterogeneity in the distribution of these structural variants according to replication timing: some had events that occurred predominantly in late-replicating regions, others had events that occurred exclusively in early-replicating regions, and in others events were distributed more evenly (Supplementary Fig. 5). Regions of active chromatin and increased gene density correlated positively with the rate of rearrangement.
The final set of structural variants used in this Article was generated by the Technical Working Group of the PCAWG Consortium and is described in the main PCAWG paper8. In brief, four variant callers were used to identify somatically acquired structural variants from matched tumour and germline whole genome sequencing data: SvABA (Broad pipeline), DELLY (DKFZ pipeline), BRASS (Sanger pipeline) and dRanger (Broad pipeline). These were merged into a final call set using a graph-based algorithm to identify overlapping breakpoint junctions across algorithms. Detailed visual inspection of structural-variant calls suggested that a simple approach of accepting all structural-variant calls made by two or more of the four algorithms gave the best trade-off between sensitivity and specificity.
This work was supported by the Wellcome Trust, Pediatric Low-Grade Astrocytoma Fund and the Fund for Innovation in Cancer Informatics. P.J.C. is a Wellcome Trust Senior Clinical Fellow (WT088340MA). We acknowledge the contributions of the many clinical networks across ICGC and TCGA, which provided samples and data to the PCAWG Consortium, and the contributions of the Technical Working Group and the Germline Working Group of the PCAWG Consortium for the collation, realignment and harmonized variant-calling of the cancer genomes used in this study. We thank the patients and their families for their participation in the individual ICGC and TCGA projects.
a, Size distribution of tandem duplications per histology group. b, Samples with more than 20 tandem duplications were grouped using hierarchical clustering according to the within-patient distribution of tandem-duplication size. Seven clusters emerged, with the size distribution of up to eight randomly chosen samples per cluster illustrated. The numbers in the top right of each panel denote the number of tandem duplications in that sample.
d0d94e66b7