Reproducible brain-wide association studies require thousands of individuals.

Magnetic resonance imaging (MRI) has transformed our understanding of the human brain through well-replicated mapping of abilities to specific structures (for example, lesion studies) and functions1-3 (for example, task functional MRI (fMRI)). Mental health research and care have yet to realize similar advances from MRI. A primary challenge has been replicating associations between inter-individual differences in brain structure or function and complex cognitive or mental health phenotypes (brain-wide association studies (BWAS)). Such BWAS have typically relied on sample sizes appropriate for classical brain mapping4 (the median neuroimaging study sample size is about 25), but potentially too small for capturing reproducible brain-behavioural phenotype associations5,6. Here we used three of the largest neuroimaging datasets currently available-with a total sample size of around 50,000 individuals-to quantify BWAS effect sizes and reproducibility as a function of sample size. BWAS associations were smaller than previously thought, resulting in statistically underpowered studies, inflated effect sizes and replication failures at typical sample sizes. As sample sizes grew into the thousands, replication rates began to improve and effect size inflation decreased. More robust BWAS effects were detected for functional MRI (versus structural), cognitive tests (versus mental health questionnaires) and multivariate methods (versus univariate). Smaller than expected brain-phenotype associations and variability across population subsamples can explain widespread BWAS replication failures. In contrast to non-BWAS approaches with larger effects (for example, lesions, interventions and within-person), BWAS reproducibility requires samples with thousands of individuals.

Comparison of individualized behavioral predictions across anatomical, diffusion and functional connectivity MRI.

A fundamental goal across the neurosciences is the characterization of relationships linking brain anatomy, functioning, and behavior. Although various MRI modalities have been developed to probe these relationships, direct comparisons of their ability to predict behavior have been lacking. Here, we compared the ability of anatomical T1, diffusion and functional MRI (fMRI) to predict behavior at an individual level. Cortical thickness, area and volume were extracted from anatomical T1 images. Diffusion Tensor Imaging (DTI) and approximate Neurite Orientation Dispersion and Density Imaging (NODDI) models were fitted to the diffusion images. The resulting metrics were projected to the Tract-Based Spatial Statistics (TBSS) skeleton. We also ran probabilistic tractography for the diffusion images, from which we extracted the stream count, average stream length, and the average of each DTI and NODDI metric across tracts connecting each pair of brain regions. Functional connectivity (FC) was extracted from both task and resting-state fMRI. Individualized prediction of a wide range of behavioral measures were performed using kernel ridge regression, linear ridge regression and elastic net regression. Consistency of the results were investigated with the Human Connectome Project (HCP) and Adolescent Brain Cognitive Development (ABCD) datasets. In both datasets, FC-based models gave the best prediction performance, regardless of regression model or behavioral measure. This was especially true for the cognitive component. Furthermore, all modalities were able to predict cognition better than other behavioral components. Combining all modalities improved prediction of cognition, but not other behavioral components. Finally, across all behaviors, combining resting and task FC yielded prediction performance similar to combining all modalities. Overall, our study suggests that in the case of healthy children and young adults, behaviorally-relevant information in T1 and diffusion features might reflect a subset of the variance captured by FC.

Integrative analysis of 111 reference human epigenomes.

The reference human genome sequence set the stage for studies of genetic variation and its association with human disease, but epigenomic studies lack a similar reference. To address this need, the NIH Roadmap Epigenomics Consortium generated the largest collection so far of human epigenomes for primary cells and tissues. Here we describe the integrative analysis of 111 reference human epigenomes generated as part of the programme, profiled for histone modification patterns, DNA accessibility, DNA methylation and RNA expression. We establish global maps of regulatory elements, define regulatory modules of coordinated activity, and their likely activators and repressors. We show that disease- and trait-associated genetic variants are enriched in tissue-specific epigenomic marks, revealing biologically relevant cell types for diverse human traits, and providing a resource for interpreting the molecular basis of human disease. Our results demonstrate the central role of epigenomic information for understanding gene regulation, cellular differentiation and human disease.

Separable Roles for a Caenorhabditis elegans RMI1 Homolog in Promoting and Antagonizing Meiotic Crossovers Ensure Faithful Chromosome Inheritance.

During the first meiotic division, crossovers (COs) between homologous chromosomes ensure their correct segregation. COs are produced by homologous recombination (HR)-mediated repair of programmed DNA double strand breaks (DSBs). As more DSBs are induced than COs, mechanisms are required to establish a regulated number of COs and to repair remaining intermediates as non-crossovers (NCOs). We show that the Caenorhabditis elegans RMI1 homolog-1 (RMH-1) functions during meiosis to promote both CO and NCO HR at appropriate chromosomal sites. RMH-1 accumulates at CO sites, dependent on known pro-CO factors, and acts to promote CO designation and enforce the CO outcome of HR-intermediate resolution. RMH-1 also localizes at NCO sites and functions in parallel with SMC-5 to antagonize excess HR-based connections between chromosomes. Moreover, RMH-1 also has a major role in channeling DSBs into an NCO HR outcome near the centers of chromosomes, thereby ensuring that COs form predominantly at off-center positions.

Brain properties predict proximity to symptom onset in sporadic Alzheimer's disease.

See Tijms and Visser (doi:10.1093/brain/awy113) for a scientific commentary on this article.Alzheimer's disease is preceded by a lengthy 'preclinical' stage spanning many years, during which subtle brain changes occur in the absence of overt cognitive symptoms. Predicting when the onset of disease symptoms will occur is an unsolved challenge in individuals with sporadic Alzheimer's disease. In individuals with autosomal dominant genetic Alzheimer's disease, the age of symptom onset is similar across generations, allowing the prediction of individual onset times with some accuracy. We extend this concept to persons with a parental history of sporadic Alzheimer's disease to test whether an individual's symptom onset age can be informed by the onset age of their affected parent, and whether this estimated onset age can be predicted using only MRI. Structural and functional MRIs were acquired from 255 ageing cognitively healthy subjects with a parental history of sporadic Alzheimer's disease from the PREVENT-AD cohort. Years to estimated symptom onset was calculated as participant age minus age of parental symptom onset. Grey matter volume was extracted from T1-weighted images and whole-brain resting state functional connectivity was evaluated using degree count. Both modalities were summarized using a 444-region cortical-subcortical atlas. The entire sample was divided into training (n = 138) and testing (n = 68) sets. Within the training set, individuals closer to or beyond their parent's symptom onset demonstrated reduced grey matter volume and altered functional connectivity, specifically in regions known to be vulnerable in Alzheimer's disease. Machine learning was used to identify a weighted set of imaging features trained to predict years to estimated symptom onset. This feature set alone significantly predicted years to estimated symptom onset in the unseen testing data. This model, using only neuroimaging features, significantly outperformed a similar model instead trained with cognitive, genetic, imaging and demographic features used in a traditional clinical setting. We next tested if these brain properties could be generalized to predict time to clinical progression in a subgroup of 26 individuals from the Alzheimer's Disease Neuroimaging Initiative, who eventually converted either to mild cognitive impairment or to Alzheimer's dementia. The feature set trained on years to estimated symptom onset in the PREVENT-AD predicted variance in time to clinical conversion in this separate longitudinal dataset. Adjusting for participant age did not impact any of the results. These findings demonstrate that years to estimated symptom onset or similar measures can be predicted from brain features and may help estimate presymptomatic disease progression in at-risk individuals.

Vitamin C induces Tet-dependent DNA demethylation and a blastocyst-like state in ES cells.

DNA methylation is a heritable epigenetic modification involved in gene silencing, imprinting, and the suppression of retrotransposons. Global DNA demethylation occurs in the early embryo and the germ line, and may be mediated by Tet (ten eleven translocation) enzymes, which convert 5-methylcytosine (5mC) to 5-hydroxymethylcytosine (5hmC). Tet enzymes have been studied extensively in mouse embryonic stem (ES) cells, which are generally cultured in the absence of vitamin C, a potential cofactor for Fe(II) 2-oxoglutarate dioxygenase enzymes such as Tet enzymes. Here we report that addition of vitamin C to mouse ES cells promotes Tet activity, leading to a rapid and global increase in 5hmC. This is followed by DNA demethylation of many gene promoters and upregulation of demethylated germline genes. Tet1 binding is enriched near the transcription start site of genes affected by vitamin C treatment. Importantly, vitamin C, but not other antioxidants, enhances the activity of recombinant Tet1 in a biochemical assay, and the vitamin-C-induced changes in 5hmC and 5mC are entirely suppressed in Tet1 and Tet2 double knockout ES cells. Vitamin C has a stronger effect on regions that gain methylation in cultured ES cells compared to blastocysts, and in vivo are methylated only after implantation. In contrast, imprinted regions and intracisternal A particle retroelements, which are resistant to demethylation in the early embryo, are resistant to vitamin-C-induced DNA demethylation. Collectively, the results of this study establish vitamin C as a direct regulator of Tet activity and DNA methylation fidelity in ES cells.

Differential responses of MET activations to MET kinase inhibitor and neutralizing antibody.

BACKGROUND: Aberrant MET tyrosine kinase signaling is known to cause cancer initiation and progression. While MET inhibitors are in clinical trials against several cancer types, the clinical efficacies are controversial and the molecular mechanisms toward sensitivity remain elusive. METHODS: With the goal to investigate the molecular basis of MET amplification (METamp) and hepatocyte growth factor (HGF) autocrine-driven tumors in response to MET tyrosine kinase inhibitors (TKI) and neutralizing antibodies, we compared cancer cells harboring METamp (MKN45 and MHCCH97H) or HGF-autocrine (JHH5 and U87) for their sensitivity and downstream biological responses to a MET-TKI (INC280) and an anti-MET monoclonal antibody (MetMab) in vitro, and for tumor inhibition in vivo. RESULTS: We find that cancer cells driven by METamp are more sensitive to INC280 than are those driven by HGF-autocrine activation. In METamp cells, INC280 induced a DNA damage response with activation of repair through the p53BP1/ATM signaling pathway. Although MetMab failed to inhibit METamp cell proliferation and tumor growth, both INC280 and MetMab reduced HGF-autocrine tumor growth. In addition, we also show that HGF stimulation promoted human HUVEC cell tube formation via the Src pathway, which was inhibited by either INC280 or MetMab. These observations suggest that in HGF-autocrine tumors, the endothelial cells are the secondary targets MET inhibitors. CONCLUSIONS: Our results demonstrate that METamp and HGF-autocrine activation favor different molecular mechanisms. While combining MET TKIs and ATM inhibitors may enhance the efficacy for treating tumors harboring METamp, a combined inhibition of MET and angiogenesis pathways may improve the therapeutic efficacy against HGF-autocrine tumors.

The clonal and mutational evolution spectrum of primary triple-negative breast cancers.

Primary triple-negative breast cancers (TNBCs), a tumour type defined by lack of oestrogen receptor, progesterone receptor and ERBB2 gene amplification, represent approximately 16% of all breast cancers. Here we show in 104 TNBC cases that at the time of diagnosis these cancers exhibit a wide and continuous spectrum of genomic evolution, with some having only a handful of coding somatic aberrations in a few pathways, whereas others contain hundreds of coding somatic mutations. High-throughput RNA sequencing (RNA-seq) revealed that only approximately 36% of mutations are expressed. Using deep re-sequencing measurements of allelic abundance for 2,414 somatic mutations, we determine for the first time-to our knowledge-in an epithelial tumour subtype, the relative abundance of clonal frequencies among cases representative of the population. We show that TNBCs vary widely in their clonal frequencies at the time of diagnosis, with the basal subtype of TNBC showing more variation than non-basal TNBC. Although p53 (also known as TP53), PIK3CA and PTEN somatic mutations seem to be clonally dominant compared to other genes, in some tumours their clonal frequencies are incompatible with founder status. Mutations in cytoskeletal, cell shape and motility proteins occurred at lower clonal frequencies, suggesting that they occurred later during tumour progression. Taken together, our results show that understanding the biology and therapeutic responses of patients with TNBC will require the determination of individual tumour clonal genotypes.

Frequent mutation of histone-modifying genes in non-Hodgkin lymphoma.

Follicular lymphoma (FL) and diffuse large B-cell lymphoma (DLBCL) are the two most common non-Hodgkin lymphomas (NHLs). Here we sequenced tumour and matched normal DNA from 13 DLBCL cases and one FL case to identify genes with mutations in B-cell NHL. We analysed RNA-seq data from these and another 113 NHLs to identify genes with candidate mutations, and then re-sequenced tumour and matched normal DNA from these cases to confirm 109 genes with multiple somatic mutations. Genes with roles in histone modification were frequent targets of somatic mutation. For example, 32% of DLBCL and 89% of FL cases had somatic mutations in MLL2, which encodes a histone methyltransferase, and 11.4% and 13.4% of DLBCL and FL cases, respectively, had mutations in MEF2B, a calcium-regulated gene that cooperates with CREBBP and EP300 in acetylating histones. Our analysis suggests a previously unappreciated disruption of chromatin biology in lymphomagenesis.

Functional DNA methylation differences between tissues, cell types, and across individuals discovered using the M&M algorithm.

DNA methylation plays key roles in diverse biological processes such as X chromosome inactivation, transposable element repression, genomic imprinting, and tissue-specific gene expression. Sequencing-based DNA methylation profiling provides an unprecedented opportunity to map and compare complete DNA methylomes. This includes one of the most widely applied technologies for measuring DNA methylation: methylated DNA immunoprecipitation followed by sequencing (MeDIP-seq), coupled with a complementary method, methylation-sensitive restriction enzyme sequencing (MRE-seq). A computational approach that integrates data from these two different but complementary assays and predicts methylation differences between samples has been unavailable. Here, we present a novel integrative statistical framework M&M (for integration of MeDIP-seq and MRE-seq) that dynamically scales, normalizes, and combines MeDIP-seq and MRE-seq data to detect differentially methylated regions. Using sample-matched whole-genome bisulfite sequencing (WGBS) as a gold standard, we demonstrate superior accuracy and reproducibility of M&M compared to existing analytical methods for MeDIP-seq data alone. M&M leverages the complementary nature of MeDIP-seq and MRE-seq data to allow rapid comparative analysis between whole methylomes at a fraction of the cost of WGBS. Comprehensive analysis of nineteen human DNA methylomes with M&M reveals distinct DNA methylation patterns among different tissue types, cell types, and individuals, potentially underscoring divergent epigenetic regulation at different scales of phenotypic diversity. We find that differential DNA methylation at enhancer elements, with concurrent changes in histone modifications and transcription factor binding, is common at the cell, tissue, and individual levels, whereas promoter methylation is more prominent in reinforcing fundamental tissue identities.

Connective tissue growth factor causes EMT-like cell fate changes in vivo and in vitro.

Connective tissue growth factor (CTGF) plays an important role in the pathogenesis of chronic fibrotic diseases. However, the mechanism by which paracrine effects of CTGF control the cell fate of neighboring epithelial cells is not known. In this study, we investigated the paracrine effects of CTGF overexpressed in fibroblasts of Col1a2-CTGF transgenic mice on epithelial cells of skin and lung. The skin and lungs of Col1a2-CTGF transgenic mice were examined for phenotypic markers of epithelial activation and differentiation and stimulation of signal transduction pathways. In addition to an expansion of the dermal compartment in Col1a2-CTGF transgenic mice, the epidermis was characterized by focal hyperplasia, and basal cells stained positive for αSMA, Snail, S100A4 and Sox9, indicating that these cells had undergone a change in their genetic program. Activation of phosphorylated p38 and phosphorylated Erk1/2 was observed in the granular and cornified layers of the skin. Lung fibrosis was associated with a marked increase in cells co-expressing epithelial and mesenchymal markers in the lesional and unaffected lung tissue of Col1a2-CTGF mice. In epithelial cells treated with TGFß, CTGF-specific siRNA-mediated knockdown suppressed Snail, Sox9, S100A4 protein levels and restored E-cadherin levels. Both adenoviral expression of CTGF in epithelial cells and treatment with recombinant CTGF induced EMT-like morphological changes and expression of α-SMA. Our in vivo and in vitro data supports the notion that CTGF expression in mesenchymal cells in the skin and lungs can cause changes in the differentiation program of adjacent epithelial cells. We speculate that these changes might contribute to fibrogenesis.

A Notch-dependent transcriptional hierarchy promotes mesenchymal transdifferentiation in the cardiac cushion.

BACKGROUND: Valvuloseptal defects are the most common congenital heart defects. Notch signaling-induced endothelial-to-mesenchymal transition (EMT) in the atrioventricular canal (AVC) cushions at murine embryonic day (E)9.5 is a required step during early valve development. Insights to the transcriptional network that is activated in endocardial cells (EC) during EMT and how these pathways direct valve maturation are lacking. RESULTS: We show that at E11.5, AVC-EC retain the ability to undergo Notch-dependent EMT when explanted on collagen. EC-Notch inhibition at E10.5 blocks expression of known mesenchymal genes in E11.5 AVC-EC. To understand the genetic network and AVC development downstream of Notch signaling beyond E9.5, we constructed Tag-Seq libraries corresponding to different cell types of the E11.5 AVC and atrium in wild-type mice and in EC-Notch inhibited mice. We identified 1,400 potential Notch targets in the AVC-EC, of which 124 are transcription factors (TF). From the 124 TFs, we constructed a transcriptional hierarchy and identify 10 upstream TFs within the network. CONCLUSIONS: We validated 4 of the upstream TFs as Notch targets that are enriched in AVC-EC. Functionally, we show these 4 TFs regulate EMT in AVC explant assays. These novel signaling pathways downstream of Notch are potentially relevant to valve development.

Complete genomic landscape of a recurring sporadic parathyroid carcinoma.

Parathyroid carcinoma is a rare endocrine malignancy with an estimated incidence of less than 1 per million population. Excessive secretion of parathyroid hormone, extremely high serum calcium level, and the deleterious effects of hypercalcaemia are the clinical manifestations of the disease. Up to 60% of patients develop multiple disease recurrences and although long-term survival is possible with palliative surgery, permanent remission is rarely achieved. Molecular drivers of sporadic parathyroid carcinoma have remained largely unknown. Previous studies, mostly based on familial cases of the disease, suggested potential roles for the tumour suppressor MEN1 and proto-oncogene RET in benign parathyroid tumourigenesis, while the tumour suppressor HRPT2 and proto-oncogene CCND1 may also act as drivers in parathyroid cancer. Here, we report the complete genomic analysis of a sporadic and recurring parathyroid carcinoma. Mutational landscapes of the primary and recurrent tumour specimens were analysed using high-throughput sequencing technologies. Such molecular profiling allowed for identification of somatic mutations never previously identified in this malignancy. These included single nucleotide point mutations in well-characterized cancer genes such as mTOR, MLL2, CDKN2C, and PIK3CA. Comparison of acquired mutations in patient-matched primary and recurrent tumours revealed loss of PIK3CA activating mutation during the evolution of the tumour from the primary to the recurrence. Structural variations leading to gene fusions and regions of copy loss and gain were identified at a single-base resolution. Loss of the short arm of chromosome 1, along with somatic missense and truncating mutations in CDKN2C and THRAP3, respectively, provides new evidence for the potential role of these genes as tumour suppressors in parathyroid cancer. The key somatic mutations identified in this study can serve as novel diagnostic markers as well as therapeutic targets.

Barnacle: detecting and characterizing tandem duplications and fusions in transcriptome assemblies.

BACKGROUND: Chimeric transcripts, including partial and internal tandem duplications (PTDs, ITDs) and gene fusions, are important in the detection, prognosis, and treatment of human cancers. RESULTS: We describe Barnacle, a production-grade analysis tool that detects such chimeras in de novo assemblies of RNA-seq data, and supports prioritizing them for review and validation by reporting the relative coverage of co-occurring chimeric and wild-type transcripts. We demonstrate applications in large-scale disease studies, by identifying PTDs in MLL, ITDs in FLT3, and reciprocal fusions between PML and RARA, in two deeply sequenced acute myeloid leukemia (AML) RNA-seq datasets. CONCLUSIONS: Our analyses of real and simulated data sets show that, with appropriate filter settings, Barnacle makes highly specific predictions for three types of chimeric transcripts that are important in a range of cancers: PTDs, ITDs, and fusions. High specificity makes manual review and validation efficient, which is necessary in large-scale disease studies. Characterizing an extended range of chimera types will help generate insights into progression, treatment, and outcomes for complex diseases.

ARID1A mutations in endometriosis-associated ovarian carcinomas.

BACKGROUND: Ovarian clear-cell and endometrioid carcinomas may arise from endometriosis, but the molecular events involved in this transformation have not been described. METHODS: We sequenced the whole transcriptomes of 18 ovarian clear-cell carcinomas and 1 ovarian clear-cell carcinoma cell line and found somatic mutations in ARID1A (the AT-rich interactive domain 1A [SWI-like] gene) in 6 of the samples. ARID1A encodes BAF250a, a key component of the SWI­SNF chromatin remodeling complex. We sequenced ARID1A in an additional 210 ovarian carcinomas and a second ovarian clear-cell carcinoma cell line and measured BAF250a expression by means of immunohistochemical analysis in an additional 455 ovarian carcinomas. RESULTS: ARID1A mutations were seen in 55 of 119 ovarian clear-cell carcinomas (46%), 10 of 33 endometrioid carcinomas (30%), and none of the 76 high-grade serous ovarian carcinomas. Seventeen carcinomas had two somatic mutations each. Loss of the BAF250a protein correlated strongly with the ovarian clear-cell carcinoma and endometrioid carcinoma subtypes and the presence of ARID1A mutations. In two patients, ARID1A mutations and loss of BAF250a expression were evident in the tumor and contiguous atypical endometriosis but not in distant endometriotic lesions. CONCLUSIONS: These data implicate ARID1A as a tumor-suppressor gene frequently disrupted in ovarian clear-cell and endometrioid carcinomas. Since ARID1A mutation and loss of BAF250a can be seen in the preneoplastic lesions, we speculate that this is an early event in the transformation of endometriosis into cancer. (Funded by the British Columbia Cancer Foundation and the Vancouver General Hospital­University of British Columbia Hospital Foundation.).

Locus co-occupancy, nucleosome positioning, and H3K4me1 regulate the functionality of FOXA2-, HNF4A-, and PDX1-bound loci in islets and liver.

The liver and pancreas share a common origin and coexpress several transcription factors. To gain insight into the transcriptional networks regulating the function of these tissues, we globally identify binding sites for FOXA2 in adult mouse islets and liver, PDX1 in islets, and HNF4A in liver. Because most eukaryotic transcription factors bind thousands of loci, many of which are thought to be inactive, methods that can discriminate functionally active binding events are essential for the interpretation of genome-wide transcription factor binding data. To develop such a method, we also generated genome-wide H3K4me1 and H3K4me3 localization data in these tissues. By analyzing our binding and histone methylation data in combination with comprehensive gene expression data, we show that H3K4me1 enrichment profiles discriminate transcription factor occupied loci into three classes: those that are functionally active, those that are poised for activation, and those that reflect pioneer-like transcription factor activity. Furthermore, we demonstrate that the regulated presence of H3K4me1-marked nucleosomes at transcription factor occupied promoters and enhancers controls their activity, implicating both tissue-specific transcription factor binding and nucleosome remodeling complex recruitment in determining tissue-specific gene expression. Finally, we apply these approaches to generate novel insights into how FOXA2, PDX1, and HNF4A cooperate to drive islet- and liver-specific gene expression.

De novo assembly and analysis of RNA-seq data.

We describe Trans-ABySS, a de novo short-read transcriptome assembly and analysis pipeline that addresses variation in local read densities by assembling read substrings with varying stringencies and then merging the resulting contigs before analysis. Analyzing 7.4 gigabases of 50-base-pair paired-end Illumina reads from an adult mouse liver poly(A) RNA library, we identified known, new and alternative structures in expressed transcripts, and achieved high sensitivity and specificity relative to reference-based assembly methods.

Distinct brain network features predict internalizing and externalizing traits in children and adults.

Internalizing and externalizing traits are two distinct classes of behaviors in psychiatry. However, whether shared or unique brain network features predict internalizing and externalizing behaviors in children and adults remain poorly understood. Using a sample of 2262 children from the Adolescent Brain Cognitive Development (ABCD) study and 752 adults from the Human Connectome Project (HCP), we show that network features predicting internalizing and externalizing behavior are, at least in part, dissociable in children, but not in adults. In ABCD children, traits within internalizing and externalizing behavioral categories are predicted by more similar network features concatenated across task and resting states than those between different categories. We did not observe this pattern in HCP adults. Distinct network features predict internalizing and externalizing behaviors in ABCD children and HCP adults. These data reveal shared and unique brain network features accounting for individual variation within broad internalizing and externalizing categories across developmental stages.

A landscape of response to drug combinations in non-small cell lung cancer.

Combination of anti-cancer drugs is broadly seen as way to overcome the often-limited efficacy of single agents. The design and testing of combinations are however very challenging. Here we present a uniquely large dataset screening over 5000 targeted agent combinations across 81 non-small cell lung cancer cell lines. Our analysis reveals a profound heterogeneity of response across the tumor models. Notably, combinations very rarely result in a strong gain in efficacy over the range of response observable with single agents. Importantly, gain of activity over single agents is more often seen when co-targeting functionally proximal genes, offering a strategy for designing more efficient combinations. Because combinatorial effect is strongly context specific, tumor specificity should be achievable. The resource provided, together with an additional validation screen sheds light on major challenges and opportunities in building efficacious combinations against cancer and provides an opportunity for training computational models for synergy prediction.