Single-molecule nanopore sequencing reveals extreme target copy number heterogeneity in arylomycin-resistant mutants.

Tandem gene amplification is a frequent and dynamic source of antibiotic resistance in bacteria. Ongoing expansions and contractions of repeat arrays during population growth are expected to manifest as cell-to-cell differences in copy number (CN). As a result, a clonal bacterial culture could comprise subpopulations of cells with different levels of antibiotic sensitivity that result from variable gene dosage. Despite the high potential for misclassification of heterogenous cell populations as either antibiotic-susceptible or fully resistant in clinical settings, and the concomitant risk of inappropriate treatment, CN distribution among cells has defied analysis. Here, we use the MinION single-molecule nanopore sequencer to uncover CN heterogeneity in clonal populations of Escherichia coli and Acinetobacter baumannii grown from single cells isolated while selecting for resistance to an optimized arylomycin, a member of a recently discovered class of Gram-negative antibiotic. We found that gene amplification of the arylomycin target, bacterial type I signal peptidase LepB, is a mechanism of unstable arylomycin resistance and demonstrate in E. coli that amplification instability is independent of RecA. This instability drives the emergence of a nonuniform distribution of lepB CN among cells with a range of 1 to at least 50 copies of lepB identified in a single clonal population. In sum, this remarkable heterogeneity, and the evolutionary plasticity it fuels, illustrates how gene amplification can enable bacterial populations to respond rapidly to novel antibiotics. This study establishes a rationale for further nanopore-sequencing studies of heterogeneous cell populations to uncover CN variability at single-molecule resolution.

Targeting PTPRK-RSPO3 colon tumours promotes differentiation and loss of stem-cell function.

Colorectal cancer remains a major unmet medical need, prompting large-scale genomics efforts in the field to identify molecular drivers for which targeted therapies might be developed. We previously reported the identification of recurrent translocations in R-spondin genes present in a subset of colorectal tumours. Here we show that targeting RSPO3 in PTPRK-RSPO3-fusion-positive human tumour xenografts inhibits tumour growth and promotes differentiation. Notably, genes expressed in the stem-cell compartment of the intestine were among those most sensitive to anti-RSPO3 treatment. This observation, combined with functional assays, suggests that a stem-cell compartment drives PTPRK-RSPO3 colorectal tumour growth and indicates that the therapeutic targeting of stem-cell properties within tumours may be a clinically relevant approach for the treatment of colorectal tumours.

CRISPR off-target analysis in genetically engineered rats and mice.

Despite widespread use of CRISPR, comprehensive data on the frequency and impact of Cas9-mediated off-targets in modified rodents are limited. Here we present deep-sequencing data from 81 genome-editing projects on mouse and rat genomes at 1,423 predicted off-target sites, 32 of which were confirmed, and show that high-fidelity Cas9 versions reduced off-target mutation rates in vivo. Using whole-genome sequencing data from ten mouse embryos, treated with a single guide RNA (sgRNA), and from their genetic parents, we found 43 off-targets, 30 of which were predicted by an adapted version of GUIDE-seq.

Evaluating the quality of the 1000 genomes project data.

BACKGROUND: Data from the 1000 Genomes project is quite often used as a reference for human genomic analysis. However, its accuracy needs to be assessed to understand the quality of predictions made using this reference. We present here an assessment of the genotyping, phasing, and imputation accuracy data in the 1000 Genomes project. We compare the phased haplotype calls from the 1000 Genomes project to experimentally phased haplotypes for 28 of the same individuals sequenced using the 10X Genomics platform. RESULTS: We observe that phasing and imputation for rare variants are unreliable, which likely reflects the limited sample size of the 1000 Genomes project data. Further, it appears that using a population specific reference panel does not improve the accuracy of imputation over using the entire 1000 Genomes data set as a reference panel. We also note that the error rates and trends depend on the choice of definition of error, and hence any error reporting needs to take these definitions into account. CONCLUSIONS: The quality of the 1000 Genomes data needs to be considered while using this database for further studies. This work presents an analysis that can be used for these assessments.

S100a4 upregulation in Pik3caH1047R;Trp53R270H;MMTV-Cre-driven mammary tumors promotes metastasis.

BACKGROUND: PIK3CA mutations are frequent in human breast cancer. Pik3caH1047R mutant expression in mouse mammary gland promotes tumorigenesis. TP53 mutations co-occur with PIK3CA mutations in human breast cancers. We previously generated a conditionally activatable Pik3caH1047R;MMTV-Cre mouse model and found a few malignant sarcomatoid (spindle cell) carcinomas that had acquired spontaneous dominant-negative Trp53 mutations. METHODS: A Pik3caH1047R;Trp53R270H;MMTV-Cre double mutant mouse breast cancer model was generated. Tumors were characterized by histology, marker analysis, transcriptional profiling, single-cell RNA-seq, and bioinformatics. Cell lines were developed from mutant tumors and used to identify and confirm genes involved in metastasis. RESULTS: We found Pik3caH1047R and Trp53R270H cooperate in driving oncogenesis in mammary glands leading to a shorter latency than either alone. Double mutant mice develop multiple histologically distinct mammary tumors, including adenocarcinoma and sarcomatoid (spindle cell) carcinoma. We found some tumors to be invasive and a few metastasized to the lung and/or the lymph node. Single-cell RNA-seq analysis of the tumors identified epithelial, stromal, myeloid, and T cell groups. Expression analysis of the metastatic tumors identified S100a4 as a top candidate gene associated with metastasis. Metastatic tumors contained a much higher percentage of epithelial-mesenchymal transition (EMT)-signature positive and S100a4-expressing cells. CRISPR/CAS9-mediated knockout of S100a4 in a metastatic tumor-derived cell line disrupted its metastatic potential indicating a role for S100a4 in metastasis. CONCLUSIONS: Pik3caH1047R;Trp53R270H;MMTV-Cre mouse provides a preclinical model to mimic a subtype of human breast cancers that carry both PIK3CA and TP53 mutations. It also allows for understanding the cooperation between the two mutant genes in tumorigenesis. Our model also provides a system to study metastasis and develop therapeutic strategies for PIK3CA/TP53 double-positive cancers. S100a4 found involved in metastasis in this model can be a potential diagnostic and therapeutic target.

Recurrent R-spondin fusions in colon cancer.

Identifying and understanding changes in cancer genomes is essential for the development of targeted therapeutics. Here we analyse systematically more than 70 pairs of primary human colon tumours by applying next-generation sequencing to characterize their exomes, transcriptomes and copy-number alterations. We have identified 36,303 protein-altering somatic changes that include several new recurrent mutations in the Wnt pathway gene TCF7L2, chromatin-remodelling genes such as TET2 and TET3 and receptor tyrosine kinases including ERBB3. Our analysis for significantly mutated cancer genes identified 23 candidates, including the cell cycle checkpoint kinase ATM. Copy-number and RNA-seq data analysis identified amplifications and corresponding overexpression of IGF2 in a subset of colon tumours. Furthermore, using RNA-seq data we identified multiple fusion transcripts including recurrent gene fusions involving R-spondin family members RSPO2 and RSPO3 that together occur in 10% of colon tumours. The RSPO fusions were mutually exclusive with APC mutations, indicating that they probably have a role in the activation of Wnt signalling and tumorigenesis. Consistent with this we show that the RSPO fusion proteins were capable of potentiating Wnt signalling. The R-spondin gene fusions and several other gene mutations identified in this study provide new potential opportunities for therapeutic intervention in colon cancer.

The BioMart community portal: an innovative alternative to large, centralized data repositories.

The BioMart Community Portal (www.biomart.org) is a community-driven effort to provide a unified interface to biomedical databases that are distributed worldwide. The portal provides access to numerous database projects supported by 30 scientific organizations. It includes over 800 different biological datasets spanning genomics, proteomics, model organisms, cancer data, ontology information and more. All resources available through the portal are independently administered and funded by their host organizations. The BioMart data federation technology provides a unified interface to all the available data. The latest version of the portal comes with many new databases that have been created by our ever-growing community. It also comes with better support and extensibility for data analysis and visualization tools. A new addition to our toolbox, the enrichment analysis tool is now accessible through graphical and web service interface. The BioMart community portal averages over one million requests per day. Building on this level of service and the wealth of information that has become available, the BioMart Community Portal has introduced a new, more scalable and cheaper alternative to the large data stores maintained by specialized organizations.

Subtype and pathway specific responses to anticancer compounds in breast cancer.

Breast cancers are comprised of molecularly distinct subtypes that may respond differently to pathway-targeted therapies now under development. Collections of breast cancer cell lines mirror many of the molecular subtypes and pathways found in tumors, suggesting that treatment of cell lines with candidate therapeutic compounds can guide identification of associations between molecular subtypes, pathways, and drug response. In a test of 77 therapeutic compounds, nearly all drugs showed differential responses across these cell lines, and approximately one third showed subtype-, pathway-, and/or genomic aberration-specific responses. These observations suggest mechanisms of response and resistance and may inform efforts to develop molecular assays that predict clinical response.

Molecular underpinnings of dedifferentiation and aggressiveness in chromophobe renal cell carcinoma.

Sarcomatoid dedifferentiation is common to multiple renal cell carcinoma (RCC) subtypes, including chromophobe RCC (ChRCC), and is associated with increased aggressiveness, resistance to targeted therapies, and heightened sensitivity to immunotherapy. To study ChRCC dedifferentiation, we performed multiregion integrated paired pathological and genomic analyses. Interestingly, ChRCC dedifferentiates not only into sarcomatoid but also into anaplastic and glandular subtypes, which are similarly associated with increased aggressiveness and metastases. Dedifferentiated ChRCC shows loss of epithelial markers, convergent gene expression, and whole genome duplication from a hypodiploid state characteristic of classic ChRCC. We identified an intermediate state with atypia and increased mitosis but preserved epithelial markers. Our data suggest that dedifferentiation is initiated by hemizygous mutation of TP53, which can be observed in differentiated areas, as well as mutation of PTEN. Notably, these mutations become homozygous with duplication of preexisting monosomes (i.e., chromosomes 17 and 10), which characterizes the transition to dedifferentiated ChRCC. Serving as potential biomarkers, dedifferentiated areas become accentuated by mTORC1 activation (phospho-S6) and p53 stabilization. Notably, dedifferentiated ChRCC share gene enrichment and pathway activation features with other sarcomatoid RCC, suggesting convergent evolutionary trajectories. This study expands our understanding of aggressive ChRCC, provides insight into molecular mechanisms of tumor progression, and informs pathologic classification and diagnostics.

Accelerated drug-resistant variant discovery with an enhanced, scalable mutagenic base editor platform.

Personalized cancer therapeutics bring directed treatment options to patients based on their tumor's genetic signature. Unfortunately, tumor genomes are remarkably adaptable, and acquired resistance through gene mutation frequently occurs. Identifying mutations that promote resistance within drug-treated patient populations can be cost, resource, and time intensive. Accordingly, base editing, enabled by Cas9-deaminase domain fusions, has emerged as a promising approach for rapid, large-scale gene variant screening in situ. Here, we adapt and optimize a conditional activation-induced cytidine deaminase (AID)-dead Cas9 (dCas9) system, which demonstrates greater heterogeneity of edits with an expanded footprint compared to the most commonly utilized cytosine base editor, BE4. In combination with a custom single guide RNA (sgRNA) library, we identify individual and compound variants in epidermal growth factor receptor (EGFR) and v-raf murine sarcoma viral oncogene homolog B1 (BRAF) that confer resistance to established EGFR inhibitors. This system and analytical pipeline provide a simple, highly scalable platform for cis or trans drug-modifying variant discovery and for uncovering valuable insights into protein structure-function relationships.

CRISPR activation screens identify the SWI/SNF ATPases as suppressors of ferroptosis.

Ferroptosis is an iron-dependent cell death mechanism characterized by the accumulation of toxic lipid peroxides and cell membrane rupture. GPX4 (glutathione peroxidase 4) prevents ferroptosis by reducing these lipid peroxides into lipid alcohols. Ferroptosis induction by GPX4 inhibition has emerged as a vulnerability of cancer cells, highlighting the need to identify ferroptosis regulators that may be exploited therapeutically. Through genome-wide CRISPR activation screens, we identify the SWI/SNF (switch/sucrose non-fermentable) ATPases BRM (SMARCA2) and BRG1 (SMARCA4) as ferroptosis suppressors. Mechanistically, they bind to and increase chromatin accessibility at NRF2 target loci, thus boosting NRF2 transcriptional output to counter lipid peroxidation and confer resistance to GPX4 inhibition. We further demonstrate that the BRM/BRG1 ferroptosis connection can be leveraged to enhance the paralog dependency of BRG1 mutant cancer cells on BRM. Our data reveal ferroptosis induction as a potential avenue for broadening the efficacy of BRM degraders/inhibitors and define a specific genetic context for exploiting GPX4 dependency.

Comparative genomics incorporating translocation renal cell carcinoma mouse model reveals molecular mechanisms of tumorigenesis.

Translocation renal cell carcinoma (tRCC) most commonly involves an ASPSCR1-TFE3 fusion, but molecular mechanisms remain elusive and animal models are lacking. Here, we show that human ASPSCR1-TFE3 driven by Pax8-Cre (a credentialed clear cell RCC driver) disrupted nephrogenesis and glomerular development, causing neonatal death, while the clear cell RCC failed driver, Sglt2-Cre, induced aggressive tRCC (as well as alveolar soft part sarcoma) with complete penetrance and short latency. However, in both contexts, ASPSCR1-TFE3 led to characteristic morphological cellular changes, loss of epithelial markers, and an epithelial-mesenchymal transition. Electron microscopy of tRCC tumors showed lysosome expansion, and functional studies revealed simultaneous activation of autophagy and mTORC1 pathways. Comparative genomic analyses encompassing an institutional human tRCC cohort (including a hitherto unreported SFPQ-TFEB fusion) and a variety of tumorgraft models (ASPSCR1-TFE3, PRCC-TFE3, SFPQ-TFE3, RBM10-TFE3, and MALAT1-TFEB) disclosed significant convergence in canonical pathways (cell cycle, lysosome, and mTORC1) and less established pathways such as Myc, E2F, and inflammation (IL-6/JAK/STAT3, interferon-γ, TLR signaling, systemic lupus, etc.). Therapeutic trials (adjusted for human drug exposures) showed antitumor activity of cabozantinib. Overall, this study provides insight into MiT/TFE-driven tumorigenesis, including the cell of origin, and characterizes diverse mouse models available for research.

Population-wide gene disruption in the murine lung epithelium via AAV-mediated delivery of CRISPR-Cas9 components.

With the aim of expediting drug target discovery and validation for respiratory diseases, we developed an optimized method for in situ somatic gene disruption in murine lung epithelial cells via AAV6-mediated CRISPR-Cas9 delivery. Efficient gene editing was observed in lung type II alveolar epithelial cells and distal airway cells following assessment of single- or dual-guide AAV vector formats, Cas9 variants, and a sequential dosing strategy with combinatorial guide RNA expression cassettes. In particular, we were able to demonstrate population-wide gene disruption within distinct epithelial cell types for separate targets in Cas9 transgenic animals, with minimal to no associated inflammation. We also observed and characterized AAV vector integration events that occurred within directed double-stranded DNA break sites in lung cells, highlighting a complicating factor with AAV-mediated delivery of DNA nucleases. Taken together, we demonstrate a uniquely effective approach for somatic engineering of the murine lung, which will greatly facilitate the modeling of disease and therapeutic intervention.

Embryonic lethality and defective mammary gland development of activator-function impaired conditional knock-in Erbb3 V943R mice.

ERBB3 is a pseudokinase domain-containing member of the ERBB family of receptor tyrosine kinases (RTKs). Following ligand binding, ERBB receptors homo- or hetero-dimerize, leading to a head-to-tail arrangement of the intracellular kinase domains, where the "receiver" kinase domain of one ERBB is activated by the "activator" domain of the other ERBB in the dimer. In ERBB3, a conserved valine at codon 943 (V943) in the kinase C-terminal domain has been shown to be important for its function as an "activator" kinase in vitro. Here we report a knock-in mouse model where we have modified the endogenous Erbb3 allele to allow for tissue-specific conditional expression of Erbb3 V943R (Erbb3 CKI-V943R ). Additionally, we generated an Erbb3 D850N (Erbb3 CKI-D850N ) conditional knock-in mouse model where the conserved aspartate in the DFG motif of the pseudokinase domain was mutated to abolish any potential residual kinase activity. While Erbb3 D850N/D850N animals developed normally, homozygous Erbb3 V943R/V943R expression during development resulted in embryonic lethality. Further, tissue specific expression of Erbb3 V943R/V943R in the mammary gland epithelium following its activation using MMTV-Cre resulted in delayed elongation of the ductal network during puberty. Single-cell RNA-seq analysis of Erbb3 V943R/V943R mammary glands showed a reduction in a specific subset of fibrinogen-producing luminal epithelial cells.

Determinants of renal cell carcinoma invasion and metastatic competence.

Metastasis is the principal cause of cancer related deaths. Tumor invasion is essential for metastatic spread. However, determinants of invasion are poorly understood. We addressed this knowledge gap by leveraging a unique attribute of kidney cancer. Renal tumors invade into large vessels forming tumor thrombi (TT) that migrate extending sometimes into the heart. Over a decade, we prospectively enrolled 83 ethnically-diverse patients undergoing surgical resection for grossly invasive tumors at UT Southwestern Kidney Cancer Program. In this study, we perform comprehensive histological analyses, integrate multi-region genomic studies, generate in vivo models, and execute functional studies to define tumor invasion and metastatic competence. We find that invasion is not always associated with the most aggressive clone. Driven by immediate early genes, invasion appears to be an opportunistic trait attained by subclones with diverse oncogenomic status in geospatial proximity to vasculature. We show that not all invasive tumors metastasize and identify determinants of metastatic competency. TT associated with metastases are characterized by higher grade, mTOR activation and a particular immune contexture. Moreover, TT grade is a better predictor of metastasis than overall tumor grade, which may have implications for clinical practice.

Prediction of epigenetically regulated genes in breast cancer cell lines.

BACKGROUND: Methylation of CpG islands within the DNA promoter regions is one mechanism that leads to aberrant gene expression in cancer. In particular, the abnormal methylation of CpG islands may silence associated genes. Therefore, using high-throughput microarrays to measure CpG island methylation will lead to better understanding of tumor pathobiology and progression, while revealing potentially new biomarkers. We have examined a recently developed high-throughput technology for measuring genome-wide methylation patterns called mTACL. Here, we propose a computational pipeline for integrating gene expression and CpG island methylation profiles to identify epigenetically regulated genes for a panel of 45 breast cancer cell lines, which is widely used in the Integrative Cancer Biology Program (ICBP). The pipeline (i) reduces the dimensionality of the methylation data, (ii) associates the reduced methylation data with gene expression data, and (iii) ranks methylation-expression associations according to their epigenetic regulation. Dimensionality reduction is performed in two steps: (i) methylation sites are grouped across the genome to identify regions of interest, and (ii) methylation profiles are clustered within each region. Associations between the clustered methylation and the gene expression data sets generate candidate matches within a fixed neighborhood around each gene. Finally, the methylation-expression associations are ranked through a logistic regression, and their significance is quantified through permutation analysis. RESULTS: Our two-step dimensionality reduction compressed 90% of the original data, reducing 137,688 methylation sites to 14,505 clusters. Methylation-expression associations produced 18,312 correspondences, which were used to further analyze epigenetic regulation. Logistic regression was used to identify 58 genes from these correspondences that showed a statistically significant negative correlation between methylation profiles and gene expression in the panel of breast cancer cell lines. Subnetwork enrichment of these genes has identified 35 common regulators with 6 or more predicted markers. In addition to identifying epigenetically regulated genes, we show evidence of differentially expressed methylation patterns between the basal and luminal subtypes. CONCLUSIONS: Our results indicate that the proposed computational protocol is a viable platform for identifying epigenetically regulated genes. Our protocol has generated a list of predictors including COL1A2, TOP2A, TFF1, and VAV3, genes whose key roles in epigenetic regulation is documented in the literature. Subnetwork enrichment of these predicted markers further suggests that epigenetic regulation of individual genes occurs in a coordinated fashion and through common regulators.

Atypical presentation and pathogenesis of a macaque lymphocryptoviral-associated B-cell lymphoma in a cynomolgus monkey.

We report the unique pathogenesis and presentation of a rapidly progressive B-cell lymphoma in a 3-year-old female cynomolgus monkey on day 50 of a 13-week toxicity study. Clinical pathology evaluation revealed a marked leukocytosis with bicytopenia. A serum protein electrophoresis was consistent with monoclonal gammopathy. The architecture of the lymph node, spleen, and thymus were variably effaced by neoplastic cells, which also infiltrated other tissues. Immunohistochemistry of the affected tissues confirmed a predominant population of CD20+, CD79a+, CD3-, CD68-, and CD34-neoplastic cells. The full data best support a diagnosis of Stage V lymphoma. Nextgen sequencing and negative prestudy serology results suggested a recent infection by macaque lymphocryptovirus (mLCV) with a unique transcriptional profile comparable with a rarely observed direct LCV infection model. This infection model might be associated with a temporary lack of an LCV antigen-specific cytotoxic T-cell adaptive immune response. Consistent with the established mechanisms of LCV-related lymphoproliferation, MYC and BCL2L11 gene expression were increased and decreased, respectively. While there was no overt immunosuppression, immunophenotyping revealed the index animal had a relatively low NK cell count, which further decreased by >50% on day 24 of the study. In addition to the temporary lack of adaptive immunity, the low NK cell counts were suggestive of an impaired innate immunity to control the virally-transformed cells and the subsequent unchecked lymphoproliferation. To our knowledge, this is the first report of a Stage V lymphoma with a unique pathogenesis in an otherwise immunocompetent cynomolgus monkey.

Elevated Serum Amino Acids Induce a Subpopulation of Alpha Cells to Initiate Pancreatic Neuroendocrine Tumor Formation.

The cellular origin of sporadic pancreatic neuroendocrine tumors (PNETs) is obscure. Hormone expression suggests that these tumors arise from glucagon-producing alpha cells or insulin-producing ß cells, but instability in hormone expression prevents linage determination. We utilize loss of hepatic glucagon receptor (GCGR) signaling to drive alpha cell hyperproliferation and tumor formation to identify a cell of origin and dissect mechanisms that drive progression. Using a combination of genetically engineered Gcgr knockout mice and GCGR-inhibiting antibodies, we show that elevated plasma amino acids drive the appearance of a proliferative population of SLC38A5+ embryonic progenitor-like alpha cells in mice. Further, we characterize tumors from patients with rare bi-allelic germline GCGR loss-of-function variants and find prominent tumor-cell-associated expression of the SLC38A5 paralog SLC7A8 as well as markers of active mTOR signaling. Thus, progenitor cells arise from adult alpha cells in response to metabolic signals and, when inductive signals are chronically present, drive tumor initiation.

Accurate assembly of the olive baboon (Papio anubis) genome using long-read and Hi-C data.

BACKGROUND: Baboons are a widely used nonhuman primate model for biomedical, evolutionary, and basic genetics research. Despite this importance, the genomic resources for baboons are limited. In particular, the current baboon reference genome Panu_3.0 is a highly fragmented, reference-guided (i.e., not fully de novo) assembly, and its poor quality inhibits our ability to conduct downstream genomic analyses. FINDINGS: Here we present a de novo genome assembly of the olive baboon (Papio anubis) that uses data from several recently developed single-molecule technologies. Our assembly, Panubis1.0, has an N50 contig size of ∼1.46 Mb (as opposed to 139 kb for Panu_3.0) and has single scaffolds that span each of the 20 autosomes and the X chromosome. CONCLUSIONS: We highlight multiple lines of evidence (including Bionano Genomics data, pedigree linkage information, and linkage disequilibrium data) suggesting that there are several large assembly errors in Panu_3.0, which have been corrected in Panubis1.0.

The Indian cobra reference genome and transcriptome enables comprehensive identification of venom toxins.

Snakebite envenoming is a serious and neglected tropical disease that kills ~100,000 people annually. High-quality, genome-enabled comprehensive characterization of toxin genes will facilitate development of effective humanized recombinant antivenom. We report a de novo near-chromosomal genome assembly of Naja naja, the Indian cobra, a highly venomous, medically important snake. Our assembly has a scaffold N50 of 223.35 Mb, with 19 scaffolds containing 95% of the genome. Of the 23,248 predicted protein-coding genes, 12,346 venom-gland-expressed genes constitute the 'venom-ome' and this included 139 genes from 33 toxin families. Among the 139 toxin genes were 19 'venom-ome-specific toxins' (VSTs) that showed venom-gland-specific expression, and these probably encode the minimal core venom effector proteins. Synthetic venom reconstituted through recombinant VST expression will aid in the rapid development of safe and effective synthetic antivenom. Additionally, our genome could serve as a reference for snake genomes, support evolutionary studies and enable venom-driven drug discovery.