VOLTA: an enVironment-aware cOntrastive ceLl represenTation leArning for histopathology.

In clinical oncology, many diagnostic tasks rely on the identification of cells in histopathology images. While supervised machine learning techniques necessitate the need for labels, providing manual cell annotations is time-consuming. In this paper, we propose a self-supervised framework (enVironment-aware cOntrastive cell represenTation learning: VOLTA) for cell representation learning in histopathology images using a technique that accounts for the cell's mutual relationship with its environment. We subject our model to extensive experiments on data collected from multiple institutions comprising over 800,000 cells and six cancer types. To showcase the potential of our proposed framework, we apply VOLTA to ovarian and endometrial cancers and demonstrate that our cell representations can be utilized to identify the known histotypes of ovarian cancer and provide insights that link histopathology and molecular subtypes of endometrial cancer. Unlike supervised models, we provide a framework that can empower discoveries without any annotation data, even in situations where sample sizes are limited.

Exploration of Germline Correlates and Risk of Immune-Related Adverse Events in Advanced Cancer Patients Treated with Immune Checkpoint Inhibitors.

Immune checkpoint inhibitors (ICIs) are increasingly used in the treatment of many tumor types, and durable responses can be observed in select populations. However, patients may exhibit significant immune-related adverse events (irAEs) that may lead to morbidity. There is limited information on whether the presence of specific germline mutations may highlight those at elevated risk of irAEs. We evaluated 117 patients with metastatic solid tumors or hematologic malignancies who underwent genomic analysis through the ongoing Personalized OncoGenomics (POG) program at BC Cancer and received an ICI during their treatment history. Charts were reviewed for irAEs. Whole genome sequencing of a fresh biopsy and matched normal specimens (blood) was performed at the time of POG enrollment. Notably, we found that MHC class I alleles in the HLA-B27 family, which have been previously associated with autoimmune conditions, were associated with grade 3 hepatitis and pneumonitis (q = 0.007) in patients treated with combination PD-1/PD-L1 and CTLA-4 inhibitors, and PD-1 inhibitors in combination with IDO-1 inhibitors. These data highlight that some patients may have a genetic predisposition to developing irAEs.

The benefit of a complete reference genome for cancer structural variant analysis.

The complexities of cancer genomes are becoming more easily interpreted due to advancements in sequencing technologies and improved bioinformatic analysis. Structural variants (SVs) represent an important subset of somatic events in tumors. While detection of SVs has been markedly improved by the development of long-read sequencing, somatic variant identification and annotation remains challenging. We hypothesized that use of a completed human reference genome (CHM13-T2T) would improve somatic SV calling. Our findings in a tumour/normal matched benchmark sample and two patient samples show that the CHM13-T2T improves SV detection and prioritization accuracy compared to GRCh38, with a notable reduction in false positive calls. We also overcame the lack of annotation resources for CHM13-T2T by lifting over CHM13-T2T-aligned reads to the GRCh38 genome, therefore combining both improved alignment and advanced annotations. In this process, we assessed the current SV benchmark set for COLO829/COLO829BL across four replicates sequenced at different centers with different long-read technologies. We discovered instability of this cell line across these replicates; 346 SVs (1.13%) were only discoverable in a single replicate. We identify 49 somatic SVs, which appear to be stable as they are consistently present across the four replicates. As such, we propose this consensus set as an updated benchmark for somatic SV calling and include both GRCh38 and CHM13-T2T coordinates in our benchmark. The benchmark is available at: 10.5281/zenodo.10819636 Our work demonstrates new approaches to optimize somatic SV prioritization in cancer with potential improvements in other genetic diseases.

Canadian COVID-19 host genetics cohort replicates known severity associations.

The HostSeq initiative recruited 10,059 Canadians infected with SARS-CoV-2 between March 2020 and March 2023, obtained clinical information on their disease experience and whole genome sequenced (WGS) their DNA. We analyzed the WGS data for genetic contributors to severe COVID-19 (considering 3,499 hospitalized cases and 4,975 non-hospitalized after quality control). We investigated the evidence for replication of loci reported by the International Host Genetics Initiative (HGI); analyzed the X chromosome; conducted rare variant gene-based analysis and polygenic risk score testing. Population stratification was adjusted for using meta-analysis across ancestry groups. We replicated two loci identified by the HGI for COVID-19 severity: the LZTFL1/SLC6A20 locus on chromosome 3 and the FOXP4 locus on chromosome 6 (the latter with a variant significant at P < 5E-8). We found novel significant associations with MRAS and WDR89 in gene-based analyses, and constructed a polygenic risk score that explained 1.01% of the variance in severe COVID-19. This study provides independent evidence confirming the robustness of previously identified COVID-19 severity loci by the HGI and identifies novel genes for further investigation.

Genomic structures and regulation patterns at HPV integration sites in cervical cancer.

Human papillomavirus (HPV) integration has been implicated in transforming HPV infection into cancer, but its genomic consequences have been difficult to study using short-read technologies. To resolve the dysregulation associated with HPV integration, we performed long-read sequencing on 63 cervical cancer genomes. We identified six categories of integration events based on HPV-human genomic structures. Of all HPV integrants, defined as two HPV-human breakpoints bridged by an HPV sequence, 24% contained variable copies of HPV between the breakpoints, a phenomenon we termed heterologous integration. Analysis of DNA methylation within and in proximity to the HPV genome at individual integration events revealed relationships between methylation status of the integrant and its orientation and structure. Dysregulation of the human epigenome and neighboring gene expression in cis with the HPV-integrated allele was observed over megabase-ranges of the genome. By elucidating the structural, epigenetic, and allele-specific impacts of HPV integration, we provide insight into the role of integrated HPV in cervical cancer.

Assembly and annotation of the black spruce genome provide insights on spruce phylogeny and evolution of stress response.

Black spruce (Picea mariana [Mill.] B.S.P.) is a dominant conifer species in the North American boreal forest that plays important ecological and economic roles. Here, we present the first genome assembly of P. mariana with a reconstructed genome size of 18.3 Gbp and NG50 scaffold length of 36.0 kbp. A total of 66,332 protein-coding sequences were predicted in silico and annotated based on sequence homology. We analyzed the evolutionary relationships between P. mariana and 5 other spruces for which complete nuclear and organelle genome sequences were available. The phylogenetic tree estimated from mitochondrial genome sequences agrees with biogeography; specifically, P. mariana was strongly supported as a sister lineage to P. glauca and 3 other taxa found in western North America, followed by the European Picea abies. We obtained mixed topologies with weaker statistical support in phylogenetic trees estimated from nuclear and chloroplast genome sequences, indicative of ancient reticulate evolution affecting these 2 genomes. Clustering of protein-coding sequences from the 6 Picea taxa and 2 Pinus species resulted in 34,776 orthogroups, 560 of which appeared to be specific to P. mariana. Analysis of these specific orthogroups and dN/dS analysis of positive selection signatures for 497 single-copy orthogroups identified gene functions mostly related to plant development and stress response. The P. mariana genome assembly and annotation provides a valuable resource for forest genetics research and applications in this broadly distributed species, especially in relation to climate adaptation.

Chemokine expression predicts T cell-inflammation and improved survival with checkpoint inhibition across solid cancers.

Immune checkpoint inhibitors (ICI) are highly effective in specific cancers where canonical markers of antitumor immunity are used for patient selection. Improved predictors of T cell-inflammation are needed to identify ICI-responsive tumor subsets in additional cancer types. We investigated associations of a 4-chemokine expression signature (c-Score: CCL4, CCL5, CXCL9, CXCL10) with metrics of antitumor immunity across tumor types. Across cancer entities from The Cancer Genome Atlas, subgroups of tumors displayed high expression of the c-Score (c-Scorehi) with increased expression of immune checkpoint (IC) genes and transcriptional hallmarks of the cancer-immunity cycle. There was an incomplete association of the c-Score with high tumor mutation burden (TMB), with only 15% of c-Scorehi tumors displaying ≥10 mutations per megabase. In a heterogeneous pan-cancer cohort of 82 patients, with advanced and previously treated solid cancers, c-Scorehi tumors had a longer median time to progression (103 versus 72 days, P = 0.012) and overall survival (382 versus 196 days, P = 0.038) following ICI therapy initiation, compared to patients with low c-Score expression. We also found c-Score stratification to outperform TMB assignment for overall survival prediction (HR = 0.42 [0.22-0.79], P = 0.008 versus HR = 0.60 [0.29-1.27], P = 0.18, respectively). Assessment of the c-Score using the TIDE and PredictIO databases, which include ICI treatment outcomes from 10 tumor types, provided further support for the c-Score as a predictive ICI therapeutic biomarker. In summary, the c-Score identifies patients with hallmarks of T cell-inflammation and potential response to ICI treatment across cancer types, which is missed by TMB assignment.

macroH2A2 antagonizes epigenetic programs of stemness in glioblastoma.

Self-renewal is a crucial property of glioblastoma cells that is enabled by the choreographed functions of chromatin regulators and transcription factors. Identifying targetable epigenetic mechanisms of self-renewal could therefore represent an important step toward developing effective treatments for this universally lethal cancer. Here we uncover an epigenetic axis of self-renewal mediated by the histone variant macroH2A2. With omics and functional assays deploying patient-derived in vitro and in vivo models, we show that macroH2A2 shapes chromatin accessibility at enhancer elements to antagonize transcriptional programs of self-renewal. macroH2A2 also sensitizes cells to small molecule-mediated cell death via activation of a viral mimicry response. Consistent with these results, our analyses of clinical cohorts indicate that high transcriptional levels of this histone variant are associated with better prognosis of high-grade glioma patients. Our results reveal a targetable epigenetic mechanism of self-renewal controlled by macroH2A2 and suggest additional treatment approaches for glioblastoma patients.

Mitogen-induced defective mitosis transforms neural progenitor cells.

BACKGROUND: Chromosome instability (CIN) with recurrent copy number alterations is a feature of many solid tumors, including glioblastoma (GBM), yet the genes that regulate cell division are rarely mutated in cancers. Here, we show that the brain-abundant mitogen, platelet-derived growth factor-A (PDGFA) fails to induce the expression of kinetochore and spindle assembly checkpoint genes leading to defective mitosis in neural progenitor cells (NPCs). METHODS: Using a recently reported in vitro model of the initiation of high-grade gliomas from murine NPCs, we investigated the immediate effects of PDGFA exposure on the nuclear and mitotic phenotypes and patterns of gene and protein expression in NPCs, a putative GBM cell of origin. RESULTS: NPCs divided abnormally in defined media containing PDGFA with P53-dependent effects. In wild-type cells, defective mitosis was associated with P53 activation and cell death, but in some null cells, defective mitosis was tolerated. Surviving cells had unstable genomes and proliferated in the presence of PDGFA accumulating random and clonal chromosomal rearrangements. The outcome of this process was a population of tumorigenic NPCs with recurrent gains and losses of chromosomal regions that were syntenic to those recurrently gained and lost in human GBM. By stimulating proliferation without setting the stage for successful mitosis, PDGFA-transformed NPCs lacking P53 function. CONCLUSIONS: Our work describes a mechanism of transformation of NPCs by a brain-associated mitogen, raising the possibility that the unique genomic architecture of GBM is an adaptation to defective mitosis that ensures the survival of affected cells.

The DACH1 gene is frequently deleted in prostate cancer, restrains prostatic intraepithelial neoplasia, decreases DNA damage repair, and predicts therapy responses.

Prostate cancer (PCa), the second leading cause of death in American men, includes distinct genetic subtypes with distinct therapeutic vulnerabilities. The DACH1 gene encodes a winged helix/Forkhead DNA-binding protein that competes for binding to FOXM1 sites. Herein, DACH1 gene deletion within the 13q21.31-q21.33 region occurs in up to 18% of human PCa and was associated with increased AR activity and poor prognosis. In prostate OncoMice, prostate-specific deletion of the Dach1 gene enhanced prostatic intraepithelial neoplasia (PIN), and was associated with increased TGFß activity and DNA damage. Reduced Dach1 increased DNA damage in response to genotoxic stresses. DACH1 was recruited to sites of DNA damage, augmenting recruitment of Ku70/Ku80. Reduced Dach1 expression was associated with increased homology directed repair and resistance to PARP inhibitors and TGFß kinase inhibitors. Reduced Dach1 expression may define a subclass of PCa that warrants specific therapies.

A generalizable machine learning framework for classifying DNA repair defects using ctDNA exomes.

Specific classes of DNA damage repair (DDR) defect can drive sensitivity to emerging therapies for metastatic prostate cancer. However, biomarker approaches based on DDR gene sequencing do not accurately predict DDR deficiency or treatment benefit. Somatic alteration signatures may identify DDR deficiency but historically require whole-genome sequencing of tumour tissue. We assembled whole-exome sequencing data for 155 high ctDNA fraction plasma cell-free DNA and matched leukocyte DNA samples from patients with metastatic prostate or bladder cancer. Labels for DDR gene alterations were established using deep targeted sequencing. Per sample mutation and copy number features were used to train XGBoost ensemble models. Naive somatic features and trinucleotide signatures were associated with specific DDR gene alterations but insufficient to resolve each class. Conversely, XGBoost-derived models showed strong performance including an area under the curve of 0.99, 0.99 and 1.00 for identifying BRCA2, CDK12, and mismatch repair deficiency in metastatic prostate cancer. Our machine learning approach re-classified several samples exhibiting genomic features inconsistent with original labels, identified a metastatic bladder cancer sample with a homozygous BRCA2 copy loss, and outperformed an existing exome-based classifier for BRCA2 deficiency. We present DARC Sign (DnA Repair Classification SIGNatures); a public machine learning tool leveraging clinically-practical liquid biopsy specimens for simultaneously identifying multiple types of metastatic prostate cancer DDR deficiencies. We posit that it will be useful for understanding differential responses to DDR-directed therapies in ongoing clinical trials and may ultimately enable prospective identification of prostate cancers with phenotypic evidence of DDR deficiency.

Homologous recombination deficiency signatures in gastrointestinal and thoracic cancers correlate with platinum therapy duration.

There is emerging evidence about the predictive role of homologous recombination deficiency (HRD), but this is less defined in gastrointestinal (GI) and thoracic malignancies. We reviewed whole genome (WGS) and transcriptomic (RNA-Seq) data from advanced GI and thoracic cancers in the Personalized OncoGenomics trial (NCT02155621) to evaluate HRD scores and single base substitution (SBS)3, which is associated with BRCA1/2 mutations and potentially predictive of defective HRD. HRD scores were calculated by sum of loss of heterozygosity, telomeric allelic imbalance, and large-scale state transitions scores. Regression analyses examined the association between HRD and time to progression on platinum (TTPp). We included 223 patients with GI (n = 154) or thoracic (n = 69) malignancies. TTPp was associated with SBS3 (p < 0.01) but not HRD score in patients with GI malignancies, whereas neither was associated with TTPp in thoracic malignancies. Tumors with gBRCA1/2 mutations and a somatic second alteration exhibited high SBS3 and HRD scores, but these signatures were also present in several tumors with germline but no somatic second alterations, suggesting silencing of the wild-type allele or BRCA1/2 haploinsufficiency. Biallelic inactivation of an HR gene, including loss of XRCC2 and BARD1, was identified in BRCA1/2 wild-type HRD tumors and these patients had prolonged response to platinum. Thoracic cases with high HRD score were associated with high RECQL5 expression (p ≤ 0.025), indicating another potential mechanism of HRD. SBS3 was more strongly associated with TTPp in patients with GI malignancies and may be complementary to using HRD and BRCA status in identifying patients who benefit from platinum therapy.

Parent-of-origin detection and chromosome-scale haplotyping using long-read DNA methylation sequencing and Strand-seq.

Hundreds of loci in human genomes have alleles that are methylated differentially according to their parent of origin. These imprinted loci generally show little variation across tissues, individuals, and populations. We show that such loci can be used to distinguish the maternal and paternal homologs for all human autosomes without the need for the parental DNA. We integrate methylation-detecting nanopore sequencing with the long-range phase information in Strand-seq data to determine the parent of origin of chromosome-length haplotypes for both DNA sequence and DNA methylation in five trios with diverse genetic backgrounds. The parent of origin was correctly inferred for all autosomes with an average mismatch error rate of 0.31% for SNVs and 1.89% for insertions or deletions (indels). Because our method can determine whether an inherited disease allele originated from the mother or the father, we predict that it will improve the diagnosis and management of many genetic diseases.

Population-size history inferences from the coho salmon (Oncorhynchus kisutch) genome.

Coho salmon (Oncorhynchus kisutch) are a culturally and economically important species that return from multiyear ocean migrations to spawn in rivers that flow to the Northern Pacific Ocean. Southern stocks of coho salmon in Canada and the United States have significantly declined over the past quarter century, and unfortunately, conservation efforts have not reversed this trend. To assist in stock management and conservation efforts, we generated a chromosome-level genome assembly. We also resequenced the genomes of 83 coho salmon across the North American range to identify nucleotide variants and understand the demographic histories of these salmon by modeling effective population size from genome-wide data. From demographic history modeling, we observed reductions in effective population sizes between 3,750 and 8,000 years ago for several northern sampling sites, which may correspond to bottleneck events during recolonization after glacial retreat.

Defining the heterogeneity of unbalanced structural variation underlying breast cancer susceptibility by nanopore genome sequencing.

Germline structural variants (SVs) are challenging to resolve by conventional genetic testing assays. Long-read sequencing has improved the global characterization of SVs, but its sensitivity at cancer susceptibility loci has not been reported. Nanopore long-read genome sequencing was performed for nineteen individuals with pathogenic copy number alterations in BRCA1, BRCA2, CHEK2 and PALB2 identified by prior clinical testing. Fourteen variants, which spanned single exons to whole genes and included a tandem duplication, were accurately represented. Defining the precise breakpoints of SVs in BRCA1 and CHEK2 revealed unforeseen allelic heterogeneity and informed the mechanisms underlying the formation of recurrent deletions. Integrating read-based and statistical phasing further helped define extended haplotypes associated with founder alleles. Long-read sequencing is a sensitive method for characterizing private, recurrent and founder SVs underlying breast cancer susceptibility. Our findings demonstrate the potential for nanopore sequencing as a powerful genetic testing assay in the hereditary cancer setting.

NCBP2 and TFRC are novel prognostic biomarkers in oral squamous cell carcinoma.

There are few prognostic biomarkers and targeted therapeutics currently in use for the clinical management of oral squamous cell carcinoma (OSCC) and patient outcomes remain poor in this disease. A majority of mutations in OSCC are loss-of-function events in tumour suppressor genes that are refractory to conventional modes of targeting. Interestingly, the chromosomal segment 3q22-3q29 is amplified in many epithelial cancers, including OSCC. We hypothesized that some of the 468 genes located on 3q22-3q29 might be drivers of oral carcinogenesis and could be exploited as potential prognostic biomarkers and therapeutic targets. Our integrative analysis of copy number variation (CNV), gene expression and clinical data from The Cancer Genome Atlas (TCGA), identified two candidate genes: NCBP2, TFRC, whose expression positively correlates with worse overall survival (OS) in HPV-negative OSCC patients. Expression of NCBP2 and TFRC is significantly higher in tumour cells compared to most normal human tissues. High NCBP2 and TFRC protein abundance is associated with worse overall, disease-specific survival, and progression-free interval in an in-house cohort of HPV-negative OSCC patients. Finally, due to a lack of evidence for the role of NCBP2 in carcinogenesis, we tested if modulating NCBP2 levels in human OSCC cell lines affected their carcinogenic behaviour. We found that NCBP2 depletion reduced OSCC cell proliferation, migration, and invasion. Differential expression analysis revealed the upregulation of several tumour-promoting genes in patients with high NCBP2 expression. We thus propose both NCBP2 and TFRC as novel prognostic and potentially therapeutic biomarkers for HPV-negative OSCC.

Phasing DNA Methylation.

Haplotyping enables the study of allele-specific events. Heterozygous variants, primarily single nucleotide variants (SNVs), enable the assignment of the paternal and maternal origin of the chromosomes and are widely employed to phase sequencing reads to their haplotype of origin. Certain long-read technologies enable the detection of both the DNA sequence and DNA modifications. These long reads and their inherent methylation information are suitable for genome-wide haplotyping and allele-specific DNA methylation analysis. Here, we describe the workflow to phase reads and DNA methylation using nanopore sequencing.

Integrative analysis of KRAS wildtype metastatic pancreatic ductal adenocarcinoma reveals mutation and expression-based similarities to cholangiocarcinoma.

Oncogenic KRAS mutations are absent in approximately 10% of patients with metastatic pancreatic ductal adenocarcinoma (mPDAC) and may represent a subgroup of mPDAC with therapeutic options beyond standard-of-care cytotoxic chemotherapy. While distinct gene fusions have been implicated in KRAS wildtype mPDAC, information regarding other types of mutations remain limited, and gene expression patterns associated with KRAS wildtype mPDAC have not been reported. Here, we leverage sequencing data from the PanGen trial to perform comprehensive characterization of the molecular landscape of KRAS wildtype mPDAC and reveal increased frequency of chr1q amplification encompassing transcription factors PROX1 and NR5A2. By leveraging data from colorectal adenocarcinoma and cholangiocarcinoma samples, we highlight similarities between cholangiocarcinoma and KRAS wildtype mPDAC involving both mutation and expression-based signatures and validate these findings using an independent dataset. These data further establish KRAS wildtype mPDAC as a unique molecular entity, with therapeutic opportunities extending beyond gene fusion events.

TMBur: a distributable tumor mutation burden approach for whole genome sequencing.

BACKGROUND: Tumor mutation burden (TMB) is a key characteristic used in a tumor-type agnostic context to inform the use of immune checkpoint inhibitors (ICI). Accurate and consistent measurement of TMB is crucial as it can significantly impact patient selection for therapy and clinical trials, with a threshold of 10 mutations/Mb commonly used as an inclusion criterion. Studies have shown that the most significant contributor to variability in mutation counts in whole genome sequence (WGS) data is differences in analysis methods, even more than differences in extraction or library construction methods. Therefore, tools for improving consistency in whole genome TMB estimation are of clinical importance. METHODS: We developed a distributable TMB analysis suite, TMBur, to address the need for genomic TMB estimate consistency in projects that span jurisdictions. TMBur is implemented in Nextflow and performs all analysis steps to generate TMB estimates directly from fastq files, incorporating somatic variant calling with Manta, Strelka2, and Mutect2, and microsatellite instability profiling with MSISensor. These tools are provided in a Singularity container downloaded by the workflow at runtime, allowing the entire workflow to be run identically on most computing platforms. To test the reproducibility of TMBur TMB estimates, we performed replicate runs on WGS data derived from the COLO829 and COLO829BL cell lines at multiple research centres. The clinical value of derived TMB estimates was then evaluated using a cohort of 90 patients with advanced, metastatic cancer that received ICIs following WGS analysis. Patients were split into groups based on a threshold of 10/Mb, and time to progression from initiation of ICIs was examined using Kaplan-Meier and cox-proportional hazards analyses. RESULTS: TMBur produced identical TMB estimates across replicates and at multiple analysis centres. The clinical utility of TMBur-derived TMB estimates were validated, with a genomic TMB ≥ 10/Mb demonstrating improved time to progression, even after correcting for differences in tumor type (HR = 0.39, p = 0.012). CONCLUSIONS: TMBur, a shareable workflow, generates consistent whole genome derived TMB estimates predictive of response to ICIs across multiple analysis centres. Reproducible TMB estimates from this approach can improve collaboration and ensure equitable treatment and clinical trial access spanning jurisdictions.