A catalog of curated breast cancer genes.

PURPOSE: Decades of research have identified multiple genetic variants associated with breast cancer etiology. However, there is no database that archives breast cancer genes and variants responsible for predisposition. We set out to build a dynamic repository of curated breast cancer genes. METHODS: A comprehensive literature search was performed in PubMed and Google Scholar, followed by data extraction and harmonization for downstream analysis. RESULTS: Using a subset of 345 studies, we cataloged 652 breast cancer-associated loci across the genome. A majority of these were present in the non-coding region (i.e., intergenic (101) and intronic (345)), whereas only 158 were located within an exon. Using the odds ratio, we identified 429 loci to increase the disease risk and 198 to confer protection against breast cancer, whereas 25 were identified to both increase disease risk and confer protection against breast cancer. Chromosomal ideogram analysis indicated that chromosomes 17 and 19 have the highest density of breast cancer loci. We manually annotated and collated breast cancer genes in which a previous association between rare-monogenic variant and breast cancer has been documented. Finally, network and functional enrichment analysis revealed that steroid metabolism and DNA repair pathways were predominant among breast cancer genes and variants. CONCLUSIONS: We have built an online interactive catalog of curated breast cancer genes ( https://cbcg.dk ). This will expedite clinical diagnostics and support the ongoing efforts in managing breast cancer etiology. Moreover, the database will serve as an essential repository when designing new breast cancer multigene panels.

A Comparison of Tools for Copy-Number Variation Detection in Germline Whole Exome and Whole Genome Sequencing Data.

Copy-number variations (CNVs) have important clinical implications for several diseases and cancers. Relevant CNVs are hard to detect because common structural variations define large parts of the human genome. CNV calling from short-read sequencing would allow single protocol full genomic profiling. We reviewed 50 popular CNV calling tools and included 11 tools for benchmarking in a reference cohort encompassing 39 whole genome sequencing (WGS) samples paired current clinical standard-SNP-array based CNV calling. Additionally, for nine samples we also performed whole exome sequencing (WES), to address the effect of sequencing protocol on CNV calling. Furthermore, we included Gold Standard reference sample NA12878, and tested 12 samples with CNVs confirmed by multiplex ligation-dependent probe amplification (MLPA). Tool performance varied greatly in the number of called CNVs and bias for CNV lengths. Some tools had near-perfect recall of CNVs from arrays for some samples, but poor precision. Several tools had better performance for NA12878, which could be a result of overfitting. We suggest combining the best tools also based on different methodologies: GATK gCNV, Lumpy, DELLY, and cn.MOPS. Reducing the total number of called variants could potentially be assisted by the use of background panels for filtering of frequently called variants.

Interpretable Autoencoders Trained on Single Cell Sequencing Data Can Transfer Directly to Data from Unseen Tissues.

Autoencoders have been used to model single-cell mRNA-sequencing data with the purpose of denoising, visualization, data simulation, and dimensionality reduction. We, and others, have shown that autoencoders can be explainable models and interpreted in terms of biology. Here, we show that such autoencoders can generalize to the extent that they can transfer directly without additional training. In practice, we can extract biological modules, denoise, and classify data correctly from an autoencoder that was trained on a different dataset and with different cells (a foreign model). We deconvoluted the biological signal encoded in the bottleneck layer of scRNA-models using saliency maps and mapped salient features to biological pathways. Biological concepts could be associated with specific nodes and interpreted in relation to biological pathways. Even in this unsupervised framework, with no prior information about cell types or labels, the specific biological pathways deduced from the model were in line with findings in previous research. It was hypothesized that autoencoders could learn and represent meaningful biology; here, we show with a systematic experiment that this is true and even transcends the training data. This means that carefully trained autoencoders can be used to assist the interpretation of new unseen data.

Genomic profiling of newly diagnosed glioblastoma patients and its potential for clinical utility - a prospective, translational study.

Glioblastoma (GBM) is an incurable brain tumor for which new treatment strategies are urgently needed. Next-generation sequencing of GBM has most often been performed retrospectively and on archival tissue from both diagnostic and relapse surgeries with limited knowledge of clinical information, including treatment given. We sought to investigate the genomic composition prospectively in treatment-naïve patients, searched for possible targetable aberrations, and investigated for prognostic and/or predictive factors. A total of 108 newly diagnosed GBM patients were included. Clinical information, progression-free survival, and overall survival (OS) were noted. Tissues were analyzed by whole-exome sequencing, single nucleotide polymorphism (SNP) and transcriptome arrays, and RNA sequencing; assessed for mutations, fusions, tumor mutational burden (TMB), and chromosomal instability (CI); and classified into GBM subgroups. Each genomic report was discussed at a multidisciplinary tumor board meeting to evaluate for matching trials. From 111 consecutive patients, 97.3% accepted inclusion in this study. Eighty-six (77%) were treated with radiation therapy/temozolomide (TMZ) and adjuvant TMZ. One NTRK2 and three FGFR3-TACC3 fusions were identified. Copy number alterations in GRB2 and SMYD4 were significantly correlated with worse median OS together with known clinical variables like age, performance status, steroid dose, and O6-methyl-guanine-DNA-methyl-transferase status. Patients with CI-median or TMB-high had significantly worse median OS compared to CI-low/high or TMB-low/median. In conclusion, performing genomic profiling at diagnosis enables evaluation of genomic-driven therapy at the first progression. Furthermore, TMB-high or CI-median patients had worse median OS, which can support the possibility of offering experimental treatment already at the first line for this group.

The Number of Signaling Pathways Altered by Driver Mutations in Chronic Lymphocytic Leukemia Impacts Disease Outcome.

PURPOSE: Investigation of signaling pathways altered by recurrent gene mutations and their clinical impact in a consecutive cohort of patients with newly diagnosed chronic lymphocytic leukemia (CLL). The heterogeneous clinical course and genetic complexity of CLL warrant improved molecular prognostication. However, the prognostic value of recurrent mutations at the time of diagnosis remains unclear. EXPERIMENTAL DESIGN: We sequenced samples from 314 consecutive, newly diagnosed patients with CLL to investigate the clinical impact of 56 recurrently mutated genes assessed by next-generation sequencing. RESULTS: Mutations were identified in 70% of patients with enrichment among IGHV unmutated cases. With 6.5 years of follow-up, 15 mutated genes investigated at the time of diagnosis demonstrated significant impact on time to first treatment (TTFT). Carrying driver mutations was associated with shorter TTFT and poor overall survival. For outcome from CLL diagnosis, the number of signaling pathways altered by driver mutations stratified patients better than the number of driver mutations. Moreover, we demonstrated gradual impact on TTFT with increasing number of altered pathways independent of CLL-IPI risk. Thus, a 25-gene, pathway-based biomarker assessing recurrent mutations refines prognostication in CLL, in particular for CLL-IPI low- and intermediate-risk patients. External validation emphasized that a broad gene panel including low burden mutations was key for the biomarker based on altered pathways. CONCLUSIONS: We propose to include the number of pathways altered by driver mutations as a biomarker together with CLL-IPI in prospective studies of CLL from time of diagnosis for incorporation into clinical care and personalized follow-up and treatment.

Deconvolution of autoencoders to learn biological regulatory modules from single cell mRNA sequencing data.

BACKGROUND: Unsupervised machine learning methods (deep learning) have shown their usefulness with noisy single cell mRNA-sequencing data (scRNA-seq), where the models generalize well, despite the zero-inflation of the data. A class of neural networks, namely autoencoders, has been useful for denoising of single cell data, imputation of missing values and dimensionality reduction. RESULTS: Here, we present a striking feature with the potential to greatly increase the usability of autoencoders: With specialized training, the autoencoder is not only able to generalize over the data, but also to tease apart biologically meaningful modules, which we found encoded in the representation layer of the network. Our model can, from scRNA-seq data, delineate biological meaningful modules that govern a dataset, as well as give information as to which modules are active in each single cell. Importantly, most of these modules can be explained by known biological functions, as provided by the Hallmark gene sets. CONCLUSIONS: We discover that tailored training of an autoencoder makes it possible to deconvolute biological modules inherent in the data, without any assumptions. By comparisons with gene signatures of canonical pathways we see that the modules are directly interpretable. The scope of this discovery has important implications, as it makes it possible to outline the drivers behind a given effect of a cell. In comparison with other dimensionality reduction methods, or supervised models for classification, our approach has the benefit of both handling well the zero-inflated nature of scRNA-seq, and validating that the model captures relevant information, by establishing a link between input and decoded data. In perspective, our model in combination with clustering methods is able to provide information about which subtype a given single cell belongs to, as well as which biological functions determine that membership.

High frequency of pathogenic germline variants within homologous recombination repair in patients with advanced cancer.

Genomic screening of cancer patients for predisposing variants is traditionally based on age at onset, family history and type of cancer. Whereas the clinical guidelines have proven efficient in identifying families exhibiting classical attributes of hereditary cancer, the frequency of patients with alternative presentations is unclear. We identified and characterized germline variants in 636 patients with advanced solid cancer using whole exome sequencing. Pathogenic and likely pathogenic germline variants among 168 genes associated with hereditary cancer were considered. These variants were identified in 17.8% of the patients and within a wide range of cancer types. In particular, patients with mesothelioma, ovarian cancer, cervical cancer, urothelial cancer, and cancer of unknown primary origin displayed high frequencies of pathogenic variants. Variants were predominantly found in DNA-repair pathways and about half were within genes involved in homologous recombination repair. Twenty-two BRCA1 and BRCA2 germline variants were identified in 12 different cancer types, of which 10 (45%) were not previously identified in these patients based on the current clinical guidelines. Loss of heterozygosity and somatic second hits were identified in several of the affected genes, supporting possible causality for cancer development. A potential treatment target based on the pathogenic germline variant could be suggested in 25 patients (4%). The study demonstrates a high frequency of pathogenic germline variants in the homologous recombination pathway in patients with advanced solid cancers. We infer that genetic screening in this group of patients may reveal high-risk families and identify patients with potential PARP inhibitor sensitive tumors.

A Gain-of-Function p53-Mutant Oncogene Promotes Cell Fate Plasticity and Myeloid Leukemia through the Pluripotency Factor FOXH1.

Mutations in the TP53 tumor suppressor gene are common in many cancer types, including the acute myeloid leukemia (AML) subtype known as complex karyotype AML (CK-AML). Here, we identify a gain-of-function (GOF) Trp53 mutation that accelerates CK-AML initiation beyond p53 loss and, surprisingly, is required for disease maintenance. The Trp53R172H mutation (TP53R175H in humans) exhibits a neomorphic function by promoting aberrant self-renewal in leukemic cells, a phenotype that is present in hematopoietic stem and progenitor cells (HSPC) even prior to their transformation. We identify FOXH1 as a critical mediator of mutant p53 function that binds to and regulates stem cell-associated genes and transcriptional programs. Our results identify a context where mutant p53 acts as a bona fide oncogene that contributes to the pathogenesis of CK-AML and suggests a common biological theme for TP53 GOF in cancer. SIGNIFICANCE: Our study demonstrates how a GOF p53 mutant can hijack an embryonic transcription factor to promote aberrant self-renewal. In this context, mutant Trp53 functions as an oncogene to both initiate and sustain myeloid leukemia and suggests a potential convergent activity of mutant Trp53 across cancer types.This article is highlighted in the In This Issue feature, p. 813.

BloodSpot: a database of healthy and malignant haematopoiesis updated with purified and single cell mRNA sequencing profiles.

BloodSpot is a gene-centric database of mRNA expression of haematopoietic cells. The web-based interface to the database includes three concomitant levels of visualization for a gene query; foremost is the expression across hematopoietic cell types, second is analysis of survival of Acute Myeloid Leukaemia patients based on gene expression, and lastly, the expression visualized in an interactive developmental tree. With the introduction of single cell data we have now also included an unbiased dimensionality reduction method to show gene expression over the continuum of haematopoiesis. The webserver includes a few select analysis functionalities, like Student's t-test, identification of correlating genes and lookup of whole genetic signatures, with the aim of making generation and testing of hypotheses quick and intuitive. The visualizations have been updated to accommodate new datatypes and the database has been largely expanded with RNA-sequencing datasets, both purified in bulk and at single cell resolution, increasing the number of single samples more than 10 fold, while keeping simplicity in presentation. The database should be of interest for any researcher within leukaemia, haematopoiesis, cellular development, or stem cells. The database is freely available at www.bloodspot.eu.

Deep targeted sequencing of TP53 in chronic lymphocytic leukemia: clinical impact at diagnosis and at time of treatment.

In chronic lymphocytic leukemia, TP53 mutations and deletion of chromosome 17p are well-characterized biomarkers associated with poor progression-free and overall survival following chemoimmunotherapy. Patients harboring low burden TP53 mutations with variant allele frequencies of 0.3-15% have been shown to have similar dismal outcome as those with high burden mutations. We here describe a highly sensitive deep targeted next-generation sequencing assay allowing for the detection of TP53 mutations as low as 0.2% variant allele frequency. Within a consecutive, single center cohort of 290 newly diagnosed patients with chronic lymphocytic leukemia, deletion of chromosome 17p was the only TP53 aberration significantly associated with shorter overall survival and treatment-free survival. We were unable to demonstrate any impact of TP53 mutations, whether high burden (variant allele frequency >10%) or low burden (variant allele frequency ≤10%), in the absence of deletion of chromosome 17p. In addition, the impact of high burden TP53 aberration (deletion of chromosome 17p and/or TP53 mutation with variant allele frequency >10%) was only evident for patients with IGHV unmutated status; no impact of TP53 aberrations on outcome was seen for patients with IGHV mutated status. In 61 patients at time of treatment, the prognostic impact of TP53 mutations over 1% variant allele frequency could be confirmed. This study furthers the identification of a clinical significant limit of detection for robust TP53 mutation analysis in chronic lymphocytic leukemia. Multicenter studies are needed for validation of ultra-sensitive TP53 mutation assays in order to define and implement a technical as well as a clinical lower limit of detection.

Circulating tumor DNA as a marker of treatment response in BRAF V600E mutated non-melanoma solid tumors.

PURPOSE: We evaluated longitudinal tracking of BRAF V600E in circulating cell-free DNA (cfDNA) as a marker of treatment response to BRAF inhibitor (BRAFi) combination therapies in non-melanoma solid tumors included in the Copenhagen Prospective Personalized Oncology (CoPPO) program. EXPERIMENTAL DESIGN: Patients with BRAF V600E-mutated tumors were treated with combination therapies including BRAFi. Quantification of mutant cfDNA from plasma was determined and correlated to clinical outcomes. Exome sequencing was performed to identify possible resistance mutations. RESULTS: Twenty-three patients had BRAF-mutated tumors out of 455 patients included in CoPPO and 17 started BRAFi combination (EGFRi/MEKi) therapy. Tumor responses were achieved in 8 out of 16 evaluable patients and the median overall- and progression-free survival (OS and PFS) was 15 and 4.8 months, respectively. Longitudinal measurements of BRAF V600E-mutant cfDNA indicated disease progression prior to radiological evaluation and a reduction in the mutant fraction of more than 50% after 4 and 12 weeks of therapy was associated with a significantly longer PFS (p=0.003 and p=0.029) and OS (p=0.029 and p=0.017). Furthermore, the baseline mutant fraction and total level of cfDNA positively correlated with tumor burden (p=0.026 and p=0.024). Finally, analysis of cfDNA at progression revealed novel mutations potentially affecting the MAPK pathway. CONCLUSION: BRAFi combination therapies showed a response rate of 50% in BRAF V600E-mutated non-melanoma tumors. The fraction of BRAF-mutant cfDNA represent a sensitive indicator for clinical outcomes with plasma collected at week 4 and 12 as crucial time points for monitoring response and disease progression.

Characterization of basal-like subtype in a Danish consecutive primary breast cancer cohort.

BACKGROUND: Transcriptome analysis enables classification of breast tumors into molecular subtypes. BRCA1/2 predisposed patients are more likely to suffer from a basal-like subtype and this group of patients displays a more distinct phenotype and genotype. Hence, in-depth characterization of this separate entity is needed. MATERIAL AND METHODS: Molecular subtyping was performed on a consecutive and unselected series of 1560 tumors from patients with primary breast cancer. Tumors were classified by the 256 gene expression signature (CIT) and associated with basic clinical characteristics and aggregated expression levels in the hallmark gene sets. RESULTS: Of the 1560 samples, 168 were classified basal-like and 120 patients were screened for BRCA1/2 mutations, resulting in 19 BRCA1/2 carriers, 95 non-carriers and six patients carried variants of unknown significance. The BRCA1/2 carriers were significantly younger and there were no carriers above 60 years of age. The tumors showed a loss in DNA-repair profile, as well as an upregulation in proliferative cancer signaling pathways. A robust molecular signature for identification of the BRCA1/2 - carriers was infeasible in the current cohort. Patients with a basal like breast cancer had the lowest median age and the largest median tumor size. They were almost exclusively diagnosed in disease stage III. CONCLUSIONS: Basal-like subtype is indeed a separate entity compared with other molecular breast cancer subtypes and the clinical course for this patient group should reflect the aggressiveness of this cancer. Taken together, patients being diagnosed with a basal-like breast cancer are in the youngest segment of breast cancer patients and are mainly diagnosed in stage III disease.

Molecular subtyping of breast cancer improves identification of both high and low risk patients.

BACKGROUND: Transcriptome analysis enables classification of breast tumors into molecular subtypes that correlate with prognosis and effect of therapy. We evaluated the clinical benefits of molecular subtyping compared to our current diagnostic practice. MATERIALS AND METHODS: Molecular subtyping was performed on a consecutive and unselected series of 524 tumors from women with primary breast cancer (n = 508). Tumors were classified by the 256 gene expression signature (CIT) and compared to conventional immunohistochemistry (IHC) procedures. RESULTS: More than 99% of tumors were eligible for molecular classification and final reports were available prior to the multidisciplinary conference. Using a prognostic standard mortality rate index (PSMRi) developed by the Danish Breast Cancer Group (DBCG) 39 patients were assigned with an intermediate risk and among these 16 (41%) were furthermore diagnosed by the multi-gene signature assigned with a luminal A tumor and consequently spared adjuvant chemotherapy. There was overall agreement between mRNA derived and IHC hormone receptor status, whereas IHC Ki67 protein proliferative index proved inaccurate, compared to the mRNA derived index. Forty-one patients with basal-like (basL) subtypes were screened for predisposing mutations regardless of clinical predisposition. Of those 17% carried pathogenic mutations. CONCLUSION: Transcriptome based subtyping of breast tumors evidently reduces the need for adjuvant chemotherapy and improves identification of women with predisposing mutations. The results imply that transcriptome profiling should become an integrated part of current breast cancer management.

Application of whole-exome sequencing to direct the specific functional testing and diagnosis of rare inherited bleeding disorders in patients from the Öresund Region, Scandinavia.

Rare inherited bleeding disorders (IBD) are a common cause of bleeding tendency. To ensure a correct diagnosis, specialized laboratory analyses are necessary. This study reports the results of an upfront diagnostic strategy using targeted whole exome sequencing. In total, 156 patients with a significant bleeding assessment tool score participated in the study, of which a third had thrombocytopenia. Eighty-seven genes specifically associated with genetic predisposition to bleeding were analysed by whole exome sequencing. Variants were classified according to the five-tier scheme. We identified 353 germline variants. Eight patients (5%) harboured a known pathogenic variant. Of the 345 previously unknown variants, computational analyses predicted 99 to be significant. Further filtration according to the Mendelian inheritance pattern, resulted in 59 variants being predicted to be clinically significant. Moreover, 34% (20/59) were assigned as novel class 4 or 5 variants upon targeted functional testing. A class 4 or 5 variant was identified in 30% of patients with thrombocytopenia (14/47) versus 11% of patients with a normal platelet count (12/109) (P < 0·01). An IBD diagnosis has a major clinical impact. The genetic investigations detailed here extricated our patients from a diagnostic conundrum, thus demonstrating that continuous optimization of the diagnostic work-up of IBD is of great benefit.