Mining genome sequencing data to identify the genomic features linked to breast cancer histopathology

Long-Lasting Complete Remission in a Patient With Metastatic Metaplastic Breast Cancer Treated With Immune Checkpoint Inhibitor and Chemotherapy: A Case Report and a Review of the Literature

Metaplastic breast cancer (MpBC) is a rare form of breast cancer known for suboptimal response to chemotherapy, high recurrence rate, poor prognosis, and limited treatment options. Recent studies have reported that MpBC has high expression of programmed death ligand 1 and tumor-infiltrating lymphocytes, indicating the potential effectiveness of immunotherapy (IO) in MpBC. In addition, several reports have demonstrated the activity of IO in MpBC. In this case report, we present a case of recurrent MpBC that achieved durable, rapid, complete remission with atezolizumab (anti-PD-L1) and nab-paclitaxel with a continued response even after discontinued therapy.

Blood-based T cell receptor anti-viral CDR3s are associated with worse overall survival for neuroblastoma

With the advent of large collections of adaptive immune receptor recombination reads representing cancer, there is the opportunity to further investigate the adaptive immune response to viruses in the cancer setting. This is a particularly important goal due to longstanding but still not well-resolved questions about viral etiologies in cancer and viral infections as comorbidities. In this report, we assessed the T cell receptor complementarity determining region-3 (CDR3) amino acid (AA) sequences, for blood-sourced TCRs from neuroblastoma (NBL) cases, for exact AA sequence matches to previously identified anti-viral TCR CDR3 AA sequences. Results indicated the presence of anti-viral TCR CDR3 AA sequences in the NBL blood samples highly significantly correlated with worse overall survival. Furthermore, the TCR CDR3 AA sequences demonstrating chemical complementarity to many cytomegalovirus antigens represented cases with a worse outcome, including cases where such CDR3s were obtained from tumor samples. Overall, these results indicate a significant need for, and provide a novel strategy for assessing viral infection complications in NBL patients.

Loss of Brca1 and Trp53 in adult mouse mammary ductal epithelium results in development of hormone receptor-positive or hormone receptor-negative tumors, depending on inactivation of Rb family proteins

BACKGROUND

Breast cancer is a heterogenous disease with several histological and molecular subtypes. Models that represent these subtypes are essential for translational research aimed at improving clinical strategy for targeted therapeutics.

METHODS

Different combinations of genetic aberrations (Brca1 and Trp53 loss, and inhibition of proteins of the Rb family) were induced in the mammary gland by injection of adenovirus expressing Cre recombinase into the mammary ducts of adult genetically engineered mice. Mammary tumors with different genetic aberrations were classified into molecular subtypes based on expression of molecular markers and RNAseq analysis. In vitro potency assays and Western blots were used to examine their drug sensitivities.

RESULTS

Induction of Brca1 and Trp53 loss in mammary ductal epithelium resulted in development of basal-like hormone receptor (HR)-negative mammary tumors. Inhibition of Rb and Trp53 loss or the combination of Rb, Trp53 and Brca1 aberrations resulted in development of luminal ductal carcinoma positive for ER, PR, and Her2 expression. HR positivity in tumors with Rb, Trp53 and Brca1 aberrations indicated that functionality of the Rb pathway rather than Brca1 status affected HR status in these models. Mammary tumor gene expression profiles recapitulated human basal-like or luminal B breast cancer signatures, but HR-positive luminal cancer models were endocrine resistant and exhibited upregulation of PI3K signaling and sensitivity to this pathway inhibition. Furthermore, both tumor subtypes were resistant to CDK4/6 inhibition.

CONCLUSIONS

Examination of molecular expression profiles and drug sensitivities of tumors indicate that these breast cancer models can be utilized as a translational platform for evaluation of targeted combinations to improve chemotherapeutic response in patients that no longer respond to hormone therapy or that are resistant to CDK4/6 inhibition.

Expression profile analysis identifies IER3 to predict overall survival and promote lymph node metastasis in tongue cancer

Background

Lymph node metastasis is one of the most important factors affecting the prognosis of tongue cancer, and the molecular mechanism regulating lymph node metastasis of tongue cancer is poorly known.

Methods

The gene expression dataset GSE2280 and The Cancer Genome Atlas (TCGA) tongue cancer dataset were downloaded. R software was used to identify the differentially expressed hallmark gene sets and individual genes between metastatic lymph node tissues and primary tongue cancer tissues, and the Kaplan-Meier method was used to evaluate the association with overall survival. The screening and validation of functional genes was performed using western blot, q-PCR, CCK-8, migration and invasion assays, and lymphangiogenesis was examined by using a tube formation assay.

Results

Thirteen common hallmark gene sets were found based on Gene Set Variation Analysis (GSVA) and then subjected to differential gene expression analysis, by which 76 deregulated genes were found. Gene coexpression network analysis and survival analysis further confirmed that IER3 was the key gene associated with the prognosis and lymph node metastasis of tongue cancer patients. Knockdown of IER3 with siRNA inhibited the proliferation, colony formation, migration and invasion of Tca-8113 cells in vitro and it also inhibited the secretion and expression of VEGF-C in these cells. The culture supernatant of Tca-8113 cells could promote lymphangiogenesis and migration of lymphatic endothelial cells, and knockdown of IER3 in Tca-8113 cells suppressed these processes.

Conclusion

Our study demonstrated that IER3 plays important roles in lymphangiogenesis regulation and prognosis in tongue cancer and might be a potential therapeutic target.

The anaphase-promoting complex: A key mitotic regulator associated with somatic mutations occurring in cancer

The anaphase-promoting complex/cyclosome (APC/C) is an E3 ubiquitin ligase that helps control chromosome separation and exit from mitosis in many different kinds of organisms, including yeast, flies, worms, and humans. This review represents a new perspective on the connection between APC/C subunit mutations and cancer. The complex nature of APC/C and limited mutation analysis of its subunits has made it difficult to determine the relationship of each subunit to cancer. In this work, cancer genomic data were examined to identify APC/C subunits with a greater than 5% alteration frequency in 11 representative cancers using the cBioPortal database. Using the Genetic Determinants of Cancer Patient Survival database, APC/C subunits were also studied and found to be significantly associated with poor patient prognosis in several cases. In comparing these two kinds of cancer genomics data to published large-scale genomic analyses looking for cancer driver genes, ANAPC1 and ANAPC3/CDC27 stood out as being represented in all three types of analyses. Seven other subunits were found to be associated both with >5% alteration frequency in certain cancers and being associated with an effect on cancer patient prognosis. The aim of this review is to provide new approaches for investigators conducting in vivo studies of APC/C subunits and cancer progression. In turn, a better understanding of these APC/C subunits and their role in different cancers will help scientists design drugs that are more precisely targeted to certain cancers, using APC/C mutation status as a biomarker.

Complete Digital Pathology for Routine Histopathology Diagnosis in a Multicenter Hospital Network

CONTEXT.—

Complete digital pathology and whole slide imaging for routine histopathology diagnosis is currently in use in few laboratories worldwide. Granada University Hospitals, Spain, which comprises 4 hospitals, adopted full digital pathology for primary histopathology diagnosis in 2016.

OBJECTIVE.—

To describe the methodology adopted and the resulting experience at Granada University Hospitals in transitioning to full digital diagnosis.

DESIGN.—

All histopathology glass slides generated for routine diagnosis were digitized at ×40 using the Philips IntelliSite Pathology Solution, which includes an ultrafast scanner and an image management system. All hematoxylin-eosin-stained preparations and immunohistochemistry and histochemistry slides were digitized. The existing sample-tracking software and image management system were integrated to allow data interchange through the Health Level 7 protocol.

RESULTS.—

Circa 160 000 specimens have been signed out using digital pathology for primary diagnosis. This comprises more than 800 000 digitized glass slides. The scanning error rate during the implementation phase was below 1.5%, and subsequent workflow optimization rendered this rate negligible. Since implementation, Granada University Hospitals pathologists have signed out 21% more cases per year on average.

CONCLUSIONS.—

Digital pathology is an adequate medium for primary histopathology diagnosis. Successful digitization relies on existing sample tracking and integration of the information technology infrastructure. Rapid and reliable scanning at ×40 equivalent was key to the transition to a fully digital workflow. Digital pathology resulted in efficiency gains in the preanalytical and analytical phases, and created the basis for computational pathology: the use of computer-assisted tools to aid diagnosis.

The Role of ZNF143 in Breast Cancer Cell Survival Through the NAD(P)H Quinone Dehydrogenase 1⁻p53⁻Beclin1 Axis Under Metabolic Stress

Autophagy is a cellular process that disrupts and uses unnecessary or malfunctioning components for cellular homeostasis. Evidence has shown a role for autophagy in tumor cell survival, but the molecular determinants that define sensitivity against autophagic regulation in cancers are not clear. Importantly, we found that breast cancer cells with low expression levels of a zinc-finger protein, ZNF143 (MCF7 sh-ZNF143), showed better survival than control cells (MCF7 sh-Control) under starvation, which was compromised with chloroquine, an autophagy inhibitor. In addition, there were more autophagic vesicles in MCF7 sh-ZNF143 cells than in MCF7 sh-Control cells, and proteins related with the autophagic process, such as Beclin1, p62, and ATGs, were altered in cells with less ZNF143. ZNF143 knockdown affected the stability of p53, which showed a dependence on MG132, a proteasome inhibitor. Data from proteome profiling in breast cancer cells with less ZNF143 suggest a role of NAD(P)H quinone dehydrogenase 1(NQO1) for p53 stability. Taken together, we showed that a subset of breast cancer cells with low expression of ZNF143 might exhibit better survival via an autophagic process by regulating the p53–Beclin1 axis, corroborating the necessity of blocking autophagy for the best therapy.

MMP7 sensitivity of mutant ECM proteins: An indicator of melanoma survival rates and T-cell infiltration

OBJECTIVE

To assess the potential impact of mutant ECM amino acids (AA) on melanoma-related matrix metalloproteinase-7 (MMP7) activity.

DESIGN AND METHODS

We applied a novel scripted algorithm, based on the MEROPS database, to reveal mutant-dependent sensitivity changes across the cancer genome atlas, melanoma dataset.

RESULTS

This approach revealed a strong bias in favor of mutant AA dependent protease sensitivity increases. Thus, melanoma specimens with relatively few mutations had only MMP7 mutant sensitive, ECM peptides. As mutations increased, melanoma specimens included mutant AA representing mostly increased sensitivity and a small but increasing number of mutant AA representing decreased MMP7 sensitivity. There was no detection of melanoma specimens with only decreases in MMP7 sensitivity. Furthermore, melanoma specimens with exclusively increased sensitivity and thereby only a few overall mutations represented reduced T-cell infiltrates and worse outcomes.

CONCLUSIONS

Overall, the results indicated that changes in MMP7 sensitivity, attributable to mutant AA, have the potential of identifying patients with distinct survival outcomes as well as patients with cancer specimen immune activity.

Mutant cytoskeletal and ECM peptides sensitive to the ST14 protease are associated with a worse outcome for glioblastoma multiforme

We previously identified a set of the most frequently mutated cytoskeleton- and extracellular matrix-related proteins (CECMPs) in numerous cancer datasets. In this report, we used a bioinformatics approach to assess the impact of amino acid (AA) substitutions on the sensitivity of CECMPs to the ST14 protease (matriptase I), a transmembrane serine protease previously implicated in cancer development. Results indicated that AA substitutions in glioblastoma multiforme (GBM) CECMPs are skewed toward increased resistance to the ST14 protease, in comparison to the wild-type peptide sequence. Furthermore, the protease resistant AA substitutions represent relatively high binding affinities to HLA class I proteins, when assessing the binding specificities using HLA class I alleles matched to the source of the mutant AA. Moreover, samples representing AA substitutions that increased protease sensitivity also represented reduced overall and disease-free survival periods for patients with glioblastoma. To assess tumor specimen immunogenicity, we identified T-cell receptor (TCR) V(D)J recombinations in GBM exome files. The overlap between ST14 protease sensitive mutant barcodes and the TCR V(D)J recombination read positive barcodes represented significantly reduced survival.

ERBB2 mutation is associated with a worse prognosis in patients with CDH1 altered invasive lobular cancer of the breast

E-cadherin (CDH1) is a glycoprotein that mediates adhesion between epithelial cells and also suppresses cancer invasion. Mutation or deletion of the CDH1 gene has been reported in 30-60% cases of invasive lobular carcinoma (ILC). However, little is known about genomic differences between ILC with and without a CDH1 alteration. Therefore, we analyzed whole genome sequencing data of 169 ILC cases from The Cancer Genome Atlas (TCGA) to address this deficiency. Our study shows that CDH1 gene was altered in 59.2% (100/169) of ILC. No significant difference was identified between CDH1-altered and -unaltered ILC cases for any of the examined demographic, clinical or pathologic characteristics, including histologic grade, tumor stage, lymph node metastases, or ER/PR/HER2 states. Seven recurrent mutations (PTEN, MUC16, ERBB2, FAT4, PCDHGA2, HERC1 and FLNC) and four chromosomal changes with recurrent copy number variation (CNV) (11q13, 17q12-21, 8p11 and 8q11) were found in ILC, which correlated with a positive or negative CDH1 alteration status, respectively. The prevalence of the most common breast cancer driver abnormalities including TP53 and PIK3CA mutations and MYC and ERBB2 amplifications showed no difference between the two groups. However, CDH1-altered ILC with an ERBB2 mutation shows a significantly worse prognosis compared to its counterparts without such a mutation. Our study suggests that CDH1-altered ILC patients with ERBB2 mutations may represent an actionable group of patients who could benefit from targeted breast cancer therapy.

Recovery of Immunoglobulin VJ Recombinations from Pancreatic Cancer Exome Files Strongly Correlates with Reduced Survival

We assessed pancreatic cancer, lymphocyte infiltrates with a computational genomics approach. We took advantage of tumor-specimen exome files available from the cancer genome atlas to mine T- and B-cell immune receptor recombinations, using highly efficient, scripted algorithms established in several previous reports. Surprisingly, the results indicated that pancreatic cancer exomes represent one of the highest level yields for immune receptor recombinations, significantly higher than two comparison cancers used in this study, head and neck and bladder cancer. In particular, pancreatic cancer exomes have very large numbers of immunoglobulin light chain recombinations, both with regard to number of samples characterized by recovery of such recombinations and with regard to numbers of recombination reads per sample. These results were consistent with B-cell biomarkers, which emphasized the Th2 nature of the pancreatic lymphocyte infiltrate. The tumor specimen exomes with B-cell immune receptor recombination reads represented a dramatically poor outcome, a result not detected with either the head and neck or bladder cancer datasets. The results presented here support the potential value of immunotherapies designed to engineer a Th2 to Th1 shift in treating certain forms of pancreatic cancer.

Identification of specific feed-forward apoptosis mechanisms and associated higher survival rates for low grade glioma and lung squamous cell carcinoma

The mechanisms of cell proliferation due to the overexpression of certain transcription factors (TFs) have been well documented in the cancer setting. However, many of these same TFs have pro-apoptotic effects, particularly when expressed or activated at high levels, a process referred to as feed-forward apoptosis (FFA). To determine whether cancers could be stratified on the basis of specific FFA signatures, RNASeq data representing samples from the cancer genome atlas were analyzed, revealing that high expression of the pro-proliferative TFs, MYC and YY1, is associated with a favorable outcome in low-grade glioma (LGG) and lung squamous cell carcinoma (LUSC), respectively. Analysis of the RNASeq data also led to the identification of specific apoptosis-effector genes whose expression levels correlate with increased survival rates, for both LGG and LUSC. Although FFA has been demonstrated as a general effect in cancer, in this report, for the first time, results identify specific TFs and their responsive effector genes that distinguish subsets of cancer samples undergoing more or less of a FFA process in a way that is associated with distinct patient survival rates.

Copy Number Profiling of MammaPrint™ Genes Reveals Association with the Prognosis of Breast Cancer Patients

Purpose

The MammaPrint™ gene signature, currently used in clinical practice, provides prognostic information regarding the recurrence and potential metastasis in breast cancer patients. However, the prognostic information of the 70 genes included can only be estimated at the RNA expression level. In this study, we investigated whether copy number information of MammaPrint™ genes at the DNA level can be used as a prognostic tool for breast cancer, as copy number variations (CNVs) are major contributors to cancer progression.

Methods

We performed CNV profiling of MammaPrint™ genes in 59 breast cancer cell lines and 650 breast cancer patients, using publicly available data in The Cancer Genome Atlas (TCGA) database. Statistical analyses including Fisher exact test, chi-square test, and Kaplan-Meier survival analyses were performed.

Results

All MammaPrint™ genes showed recurrent CNVs, particularly in TCGA cohort. CNVs of 32 and 36 genes showed significant associations with progesterone receptor and estrogen rector, respectively. No genes showed a significant association with human epidermal growth factor receptor 2 status and lymph node status. In addition, only six genes were associated with tumor stages. RFC4, HRASLS, NMU, GPR126, SCUBE2, C20orf46, and EBF4 were associated with reduced survival and RASSF7 and ESM1 were associated with reduced disease-free survival.

Conclusion

Based on these findings, a concordance of CNV-based genomic rearrangement with expression profiling of these genes and their putative roles in disease tumorigenesis was established. The results suggested that the CNV profiles of the MammaPrint™ genes can be used to predict the prognosis of breast cancer patients. In addition, this approach may lead to the development of new cancer biomarkers at the DNA level.

Dramatic response of metaplastic breast cancer to chemo-immunotherapy

Frequent overexpression of programmed death-ligand 1 has recently been demonstrated in metaplastic breast cancer, which is a rare breast cancer subtype with limited treatment options. This report describes the clinical course of a patient with metastatic metaplastic breast cancer who had a remarkable response to anti-programmed death-1 therapy with pembrolizumab in combination with nab-paclitaxel. Tissue correlates are presented including tumor-infiltrating lymphocytes and high-programmed death-ligand 1 expression in the tumor.

Genome and transcriptome delineation of two major oncogenic pathways governing invasive ductal breast cancer development

Invasive ductal carcinoma (IDC) is a major histo-morphologic type of breast cancer. Histological grading (HG) of IDC is widely adopted by oncologists as a prognostic factor. However, HG evaluation is highly subjective with only 50%-85% inter-observer agreements. Specifically, the subjectivity in the assignment of the intermediate grade (histologic grade 2, HG2) breast cancers (comprising ~50% of IDC cases) results in uncertain disease outcome prediction and sub-optimal systemic therapy. Despite several attempts to identify the mechanisms underlying the HG classification, their molecular bases are poorly understood.We performed integrative bioinformatics analysis of TCGA and several other cohorts (total 1246 patients). We identified a 22-gene tumor aggressiveness grading classifier (22g-TAG) that reflects global bifurcation in the IDC transcriptomes and reclassified patients with HG2 tumors into two genetically and clinically distinct subclasses: histological grade 1-like (HG1-like) and histological grade 3-like (HG3-like). The expression profiles and clinical outcomes of these subclasses were similar to the HG1 and HG3 tumors, respectively. We further reclassified IDC into low genetic grade (LGG = HG1+HG1-like) and high genetic grade (HGG = HG3-like+HG3) subclasses. For the HG1-like and HG3-like IDCs we found subclass-specific DNA alterations, somatic mutations, oncogenic pathways, cell cycle/mitosis and stem cell-like expression signatures that discriminate between these tumors. We found similar molecular patterns in the LGG and HGG tumor classes respectively.Our results suggest the existence of two genetically-predefined IDC classes, LGG and HGG, driven by distinct oncogenic pathways. They provide novel prognostic and therapeutic biomarkers and could open unique opportunities for personalized systemic therapies of IDC patients.

The Landscape of Whole-genome Alterations and Pathologic Features in Genitourinary Malignancies: An Analysis of the Cancer Genome Atlas

BACKGROUND

The accumulation of somatic genetic alterations drives carcinogenesis. Little is known, however, about how the level of genetic alteration across an entire cancer genome affects tumor grade, stage or survival.

OBJECTIVE

To investigate the influence of somatic mutation count (MC) and copy number variation (CNV) on pathologic and oncologic outcomes in patients with genitourinary malignancies in The Cancer Genome Atlas (TCGA).

DESIGN, SETTING, AND PARTICIPANTS

TCGA data sets for adrenocortical carcinoma (ACC), bladder urothelial carcinoma (BLCA), chromophobe renal cell carcinoma (RCC; KICH), clear cell RCC (KIRC), papillary RCC (KIRP), pheochromocytoma and paraganglioma (PCPG), prostate adenocarcinoma (PRAD), and testis germ cell tumor (TGCT) were accessed via cBioportal.

OUTCOME MEASUREMENTS AND STATISTICAL ANALYSIS

Median MC and CNV were compared among and within each tumor type. Patients were stratified by grade and stage, and differences in MC and CNV were compared. Correlation of MC and CNV with overall survival (OS) and recurrence-free survival (RFS) was analyzed when these data were available.

RESULTS AND LIMITATIONS

Among the tumor types analyzed, BLCA had the highest MC at 167, followed by ACC (89), KIRP (71), TGCT (55), KIRC (45), PRAD (34), PCPG (14), and KICH (12). The tumor type with the highest fraction of the genome with CNV was KICH (0.94), followed by ACC (0.58), TGCT (0.41), BLCA (0.29), KIRP (0.15), PCPG (0.13), KIRC (0.12), and PRAD (0.06). MC was associated with higher T stage in ACC, N stage in KIRC, M stage in ACC, grade in BLCA, and primary Gleason score in PRAD, and was associated with OS and RFS in KICH. CNV was associated with higher N stage in PRAD, grade in KIRC, and Gleason grade in PRAD. In addition, higher CNV was independently associated with inferior RFS for KIRC, as well as inferior OS and RFS for KIRP.

CONCLUSIONS

MC and CNV vary greatly among tumor types.

PATIENT SUMMARY

Cancers with higher levels of genomic alterations are associated with worse pathologic features and survival. The degree of genomic alterations could serve as a useful marker of disease aggressiveness.

A microscopic landscape of the invasive breast cancer genome

Histologic grade is one of the most important microscopic features used to predict the prognosis of invasive breast cancer and may serve as a marker for studying cancer driving genomic abnormalities in vivo. We analyzed whole genome sequencing data from 680 cases of TCGA invasive ductal carcinomas of the breast and correlated them to corresponding pathology information. Ten genetic abnormalities were found to be statistically associated with histologic grade, including three most prevalent cancer driver events, TP53 and PIK3CA mutations and MYC amplification. A distinct genetic interaction among these genomic abnormalities was revealed as measured by the histologic grading score. While TP53 mutation and MYC amplification were synergistic in promoting tumor progression, PIK3CA mutation was found to have alleviated the oncogenic effect of either the TP53 mutation or MYC amplification, and was associated with a significant reduction in mitotic activity in TP53 mutated and/or MYC amplified breast cancer. Furthermore, we discovered that different types of genetic abnormalities (mutation versus amplification) within the same cancer driver gene (PIK3CA or GATA3) were associated with opposite histologic changes in invasive breast cancer. In conclusion, our study suggests that histologic grade may serve as a biomarker to define cancer driving genetic events in vivo.

Integration of Multi-Modal Biomedical Data to Predict Cancer Grade and Patient Survival

The Big Data era in Biomedical research has resulted in large-cohort data repositories such as The Cancer Genome Atlas (TCGA). These repositories routinely contain hundreds of matched patient samples for genomic, proteomic, imaging, and clinical data modalities, enabling holistic and multi-modal integrative analysis of human disease. Using TCGA renal and ovarian cancer data, we conducted a novel investigation of multi-modal data integration by combining histopathological image and RNA-seq data. We compared the performances of two integrative prediction methods: majority vote and stacked generalization. Results indicate that integration of multiple data modalities improves prediction of cancer grade and outcome. Specifically, stacked generalization, a method that integrates multiple data modalities to produce a single prediction result, outperforms both single-data-modality prediction and majority vote. Moreover, stacked generalization reveals the contribution of each data modality (and specific features within each data modality) to the final prediction result and may provide biological insights to explain prediction performance.

A novel gammaretroviral shuttle vector insertional mutagenesis screen identifies SHARPIN as a breast cancer metastasis gene and prognostic biomarker

Breast cancer (BC) is the second leading cause of malignancy among U.S. women. Metastasis results in a poor prognosis and increased mortality, but the molecular mechanisms by which metastatic tumors occur are not well understood. Identifying the genes that drive the metastatic process could provide targets for improved therapy and biomarkers to improve BC patient outcomes. Using a forward mutagenesis screen, BC cells mutagenized with a replication-incompetent gammaretroviral vector (γRV) were xenotransplanted into the mammary fat pad of immunodeficient mice. In this approach the vector provirus dysregulates nearby genes, providing a selective advantage to transduced cells to form metastases. Metastatic tumors were analyzed for proviral integration sites to identify nearby candidate metastasis genes. The γRV has a transgene cassette that allows for rescue in bacteria and rapid identification of vector integration sites. Using this approach, we identified the previously described metastasis gene WWTR1 (TAZ), and three other novel candidate metastasis genes including SHARPIN. SHARPIN was independently validated in vivo as a BC metastasis gene. Analysis of patient data showed that SHARPIN expression predicts metastasis-free survival after adjuvant therapy. Our approach has broad potential to identify genes involved in oncogenic processes for BC and other cancers. We show here it can identify both known (WWTR1) and novel (SHARPIN) BC metastasis genes.