A deep learning model of tumor cell architecture elucidates response and resistance to CDK4/6 inhibitors.

Cyclin-dependent kinase 4 and 6 inhibitors (CDK4/6is) have revolutionized breast cancer therapy. However, <50% of patients have an objective response, and nearly all patients develop resistance during therapy. To elucidate the underlying mechanisms, we constructed an interpretable deep learning model of the response to palbociclib, a CDK4/6i, based on a reference map of multiprotein assemblies in cancer. The model identifies eight core assemblies that integrate rare and common alterations across 90 genes to stratify palbociclib-sensitive versus palbociclib-resistant cell lines. Predictions translate to patients and patient-derived xenografts, whereas single-gene biomarkers do not. Most predictive assemblies can be shown by CRISPR-Cas9 genetic disruption to regulate the CDK4/6i response. Validated assemblies relate to cell-cycle control, growth factor signaling and a histone regulatory complex that we show promotes S-phase entry through the activation of the histone modifiers KAT6A and TBL1XR1 and the transcription factor RUNX1. This study enables an integrated assessment of how a tumor's genetic profile modulates CDK4/6i resistance.

Multimodal perturbation analyses of cyclin-dependent kinases reveal a network of synthetic lethalities associated with cell-cycle regulation and transcriptional regulation.

Cell-cycle control is accomplished by cyclin-dependent kinases (CDKs), motivating extensive research into CDK targeting small-molecule drugs as cancer therapeutics. Here we use combinatorial CRISPR/Cas9 perturbations to uncover an extensive network of functional interdependencies among CDKs and related factors, identifying 43 synthetic-lethal and 12 synergistic interactions. We dissect CDK perturbations using single-cell RNAseq, for which we develop a novel computational framework to precisely quantify cell-cycle effects and diverse cell states orchestrated by specific CDKs. While pairwise disruption of CDK4/6 is synthetic-lethal, only CDK6 is required for normal cell-cycle progression and transcriptional activation. Multiple CDKs (CDK1/7/9/12) are synthetic-lethal in combination with PRMT5, independent of cell-cycle control. In-depth analysis of mRNA expression and splicing patterns provides multiple lines of evidence that the CDK-PRMT5 dependency is due to aberrant transcriptional regulation resulting in premature termination. These inter-dependencies translate to drug-drug synergies, with therapeutic implications in cancer and other diseases.

Uncovering Tumorigenesis Circuitry with Combinatorial CRISPR.

Oncogenesis relies on the alteration of multiple driver genes, but precisely which groups of alterations lead to cancer is not well understood. To chart these combinations, Zhao and colleagues use the CRISPR-Cas9 system to knockout all pairwise combinations among 52 tumor suppressor genes, with the goal of identifying groups of alterations that collaborate to promote cell growth. Interaction screens are performed across multiple models of tumorigenesis in cell cultures and mice, revealing clear cooperation among NF2, PTEN, and TP53 in multiple models. These and other strongly synergistic interactions are characterized further by single-cell transcriptomic profiling. This methodology presents a scalable approach to move beyond single-gene drivers to map the complex gene networks that give rise to tumorigenesis.See related article by Zhao et al., p. 6090.

Few-shot learning creates predictive models of drug response that translate from high-throughput screens to individual patients.

Cell-line screens create expansive datasets for learning predictive markers of drug response, but these models do not readily translate to the clinic with its diverse contexts and limited data. In the present study, we apply a recently developed technique, few-shot machine learning, to train a versatile neural network model in cell lines that can be tuned to new contexts using few additional samples. The model quickly adapts when switching among different tissue types and in moving from cell-line models to clinical contexts, including patient-derived tumor cells and patient-derived xenografts. It can also be interpreted to identify the molecular features most important to a drug response, highlighting critical roles for RB1 and SMAD4 in the response to CDK inhibition and RNF8 and CHD4 in the response to ATM inhibition. The few-shot learning framework provides a bridge from the many samples surveyed in high-throughput screens (n-of-many) to the distinctive contexts of individual patients (n-of-one).

Predicting Drug Response and Synergy Using a Deep Learning Model of Human Cancer Cells.

Most drugs entering clinical trials fail, often related to an incomplete understanding of the mechanisms governing drug response. Machine learning techniques hold immense promise for better drug response predictions, but most have not reached clinical practice due to their lack of interpretability and their focus on monotherapies. We address these challenges by developing DrugCell, an interpretable deep learning model of human cancer cells trained on the responses of 1,235 tumor cell lines to 684 drugs. Tumor genotypes induce states in cellular subsystems that are integrated with drug structure to predict response to therapy and, simultaneously, learn biological mechanisms underlying the drug response. DrugCell predictions are accurate in cell lines and also stratify clinical outcomes. Analysis of DrugCell mechanisms leads directly to the design of synergistic drug combinations, which we validate systematically by combinatorial CRISPR, drug-drug screening in vitro, and patient-derived xenografts. DrugCell provides a blueprint for constructing interpretable models for predictive medicine.

Consensus statements on the clinical uses of pregabalin for Hong Kong.

To facilitate the understanding of pregabalin and optimize its clinical usage in Hong Kong, an expert panel (11 psychiatrists, one family physician and one anesthesiologist) experienced in treating anxiety and somatic symptoms was invited to establish a set of consensus statements based on several discussion areas. A non-systematic literature search for relevant articles was conducted. The panelists addressed the discussion areas by sharing their clinical experience and available literature in a couple of meetings. At the last meeting, consensus statements derived from the proceedings were discussed and finalized. A total of 11 statements were ultimately accepted by panel voting based on their practicability of recommendation in Hong Kong. These statements are aimed to act as a practical reference for local clinicians when they consider prescribing pregabalin in different clinical situations.

Quantitative Translation of Dog-to-Human Aging by Conserved Remodeling of the DNA Methylome.

All mammals progress through similar physiological stages throughout life, from early development to puberty, aging, and death. Yet, the extent to which this conserved physiology reflects underlying genomic events is unclear. Here, we map the common methylation changes experienced by mammalian genomes as they age, focusing on comparison of humans with dogs, an emerging model of aging. Using oligo-capture sequencing, we characterize methylomes of 104 Labrador retrievers spanning a 16-year age range, achieving >150× coverage within mammalian syntenic blocks. Comparison with human methylomes reveals a nonlinear relationship that translates dog-to-human years and aligns the timing of major physiological milestones between the two species, with extension to mice. Conserved changes center on developmental gene networks, which are sufficient to translate age and the effects of anti-aging interventions across multiple mammals. These results establish methylation not only as a diagnostic age readout but also as a cross-species translator of physiological aging milestones.

A Fast and Flexible Framework for Network-Assisted Genomic Association.

We present an accessible, fast, and customizable network propagation system for pathway boosting and interpretation of genome-wide association studies. This system-NAGA (Network Assisted Genomic Association)-taps the NDEx biological network resource to gain access to thousands of protein networks and select those most relevant and performative for a specific association study. The method works efficiently, completing genome-wide analysis in under 5 minutes on a modern laptop computer. We show that NAGA recovers many known disease genes from analysis of schizophrenia genetic data, and it substantially boosts associations with previously unappreciated genes such as amyloid beta precursor. On this and seven other gene-disease association tasks, NAGA outperforms conventional approaches in recovery of known disease genes and replicability of results. Protein interactions associated with disease are visualized and annotated in Cytoscape, which, in addition to standard programmatic interfaces, allows for downstream analysis.

Strategies for Network GWAS Evaluated Using Classroom Crowd Science.

Biological networks can substantially boost power to identify disease genes in genome-wide association studies. To explore different network GWAS methods, we challenged students of a UC San Diego graduate level bioinformatics course, Network Biology and Biomedicine, to explore and improve such algorithms during a four-week-long classroom competition. Here, we report the many creative solutions and share our experiences in conducting classroom crowd science as both a research and pedagogical tool.

DDOT: A Swiss Army Knife for Investigating Data-Driven Biological Ontologies.

Systems biology requires not only genome-scale data but also methods to integrate these data into interpretable models. Previously, we developed approaches that organize omics data into a structured hierarchy of cellular components and pathways, called a "data-driven ontology." Such hierarchies recapitulate known cellular subsystems and discover new ones. To broadly facilitate this type of modeling, we report the development of a software library called the Data-Driven Ontology Toolkit (DDOT), consisting of a Python package (https://github.com/idekerlab/ddot) to assemble and analyze ontologies and a web application (http://hiview.ucsd.edu) to visualize them. Using DDOT, we programmatically assemble a compendium of ontologies for 652 diseases by integrating gene-disease mappings with a gene similarity network derived from omics data. For example, the ontology for Fanconi anemia describes known and novel disease mechanisms in its hierarchy of 194 genes and 74 subsystems. DDOT provides an easy interface to share ontologies online at the Network Data Exchange.

Using deep learning to model the hierarchical structure and function of a cell.

Although artificial neural networks are powerful classifiers, their internal structures are hard to interpret. In the life sciences, extensive knowledge of cell biology provides an opportunity to design visible neural networks (VNNs) that couple the model's inner workings to those of real systems. Here we develop DCell, a VNN embedded in the hierarchical structure of 2,526 subsystems comprising a eukaryotic cell (http://d-cell.ucsd.edu/). Trained on several million genotypes, DCell simulates cellular growth nearly as accurately as laboratory observations. During simulation, genotypes induce patterns of subsystem activities, enabling in silico investigations of the molecular mechanisms underlying genotype-phenotype associations. These mechanisms can be validated, and many are unexpected; some are governed by Boolean logic. Cumulatively, 80% of the importance for growth prediction is captured by 484 subsystems (21%), reflecting the emergence of a complex phenotype. DCell provides a foundation for decoding the genetics of disease, drug resistance and synthetic life.

Mental morbidities and chronic fatigue in severe acute respiratory syndrome survivors: long-term follow-up.

BACKGROUND: Short-term follow-up studies of severe acute respiratory syndrome (SARS) survivors suggested that their physical conditions continuously improved in the first year but that their mental health did not. We investigated long-term psychiatric morbidities and chronic fatigue among SARS survivors. METHODS: All SARS survivors from the hospitals of a local region in Hong Kong were assessed by a constellation of psychometric questionnaires and a semistructured clinical interview for the Diagnostic and Statistical Manual of Mental Disorders (Fourth Edition) to determine the presence of psychiatric disorders and chronic fatigue problems. RESULTS: Of 369 SARS survivors, 233 (63.1%) participated in the study (mean period of time after SARS, 41.3 months). Over 40% of the respondents had active psychiatric illnesses, 40.3% reported a chronic fatigue problem, and 27.1% met the modified 1994 Centers for Disease Control and Prevention criteria for chronic fatigue syndrome. Logistic regression analysis suggested that being a health care worker at the time of SARS infection (odds ratio [OR], 3.24; 95% confidence interval [CI], 1.12- 9.39; P = .03), being unemployed at follow-up (OR, 4.71; 95% CI, 1.50-14.78; P = .008), having a perception of social stigmatization (OR, 3.03; 95% CI, 1.20-7.60; P = .02), and having applied to the SARS survivors' fund (OR, 2.92; 95% CI, 1.18-7.22; P = .02) were associated with an increased risk of psychiatric morbidities at follow-up, whereas application to the SARS survivors' fund (OR, 2.64; 95% CI, 1.07-6.51; P = .04) was associated with increased risk of chronic fatigue problems. CONCLUSIONS: Psychiatric morbidities and chronic fatigue persisted and continued to be clinically significant among the survivors at the 4-year follow-up. Optimization of the treatment of mental health morbidities by a multidisciplinary approach with a view for long-term rehabilitation, especially targeting psychiatric and fatigue problems and functional and occupational rehabilitation, would be needed.

Sleepwalking in psychiatric patients: comparison of childhood and adult onset.

OBJECTIVES: In contrast to the 'benign and self-limiting nature' of childhood sleepwalking, some population and case studies have suggested that adult sleepwalking is more likely to be associated with psychopathology and psychotropic medications. There is a paucity, however, of systematic study in adult psychiatric populations, and the aim of the present study was therefore to compare the impact of psychopathology and medication usage on sleepwalking with reference to age of onset. METHODS: Clinical characteristics, sleep symptoms, psychiatric diagnosis and psychotropic usage in 66 childhood- and adult-onset sleepwalkers as identified from a psychiatric clinic, were studied. RESULTS: There was a higher proportion of adult-onset sleepwalking in the psychiatric population. In comparison with childhood-onset sleepwalkers, adult-onset sleepwalkers had higher peak frequency of attacks and a high comorbidity with sleep-related eating features. Factors including frequent insomnia (odds ratio (OR) = 5.39, 95% confidence interval (CI) = 1.58-18.40, p = 0.007) and lifetime usage of regular zolpidem (OR = 5.58, 95%CI = 1.65-18.84, p < 0.006) were associated with a higher risk of adult-onset sleepwalking. CONCLUSIONS: Adult-onset sleepwalking in a psychiatric sample has unique clinical characteristics and specific risk factors. These patients were more likely to present with sleep-related eating features, comorbid insomnia, had and lifetime usage of non-benzodiazepine hypnotics, especially zolpidem. A heightened awareness of the presence of sleepwalking and their associated risk factors among the adult psychiatric population is needed.

Parasomnia among psychiatric outpatients: a clinical, epidemiologic, cross-sectional study.

OBJECTIVE: Epidemiologic studies from general population and clinical case series suggest association of parasomnias with mental illnesses and psychotropic medications. This cross-sectional study aimed at determining the prevalence rate of sleepwalking, sleep-related eating disorder (SRED), rapid eye movement sleep behavior-like disorder (RSBD-like disorder), and sleep-related injury (SRI) and their associated factors in an adult psychiatric outpatient clinic. METHOD: Subjects aged 18 to 65 years who were attending an outpatient clinic in Hong Kong from May 2006 through June 2006 were included in this cross-sectional study. A 3-phase design was employed, including a structured questionnaire on parasomnias, followed by clinical interviews of both questionnaire-positive and -negative groups, and polysomnography for subjects having active parasomnias in recent 1 year. In addition, the principal psychiatric diagnoses, medical illnesses, and detailed drug history over recent 1 year were retrieved from the computerized records. RESULTS: Twelve hundred thirty-five subjects completed the phase 1 interview. The estimated prevalence of the lifetime diagnoses of sleepwalking, SRED, SRI, sleep violence, and RSBD-like disorder were 8.5%, 4.0%, 21.0%, 3.6%, and 5.8%, respectively, while the 1-year prevalence of these conditions were 2.9%, 2.4%, 8.8%, 2.5%, and 3.8%, respectively. These conditions were associated with depression and a constellation of sleep disturbances. Specific combinations of psychotropics were found to pose risk in particular parasomnias: sedative antidepressants and nonbenzodiazapine hypnotics in sleepwalking, regular zolpidem and antidepressants in SRED, and selective serotonin reuptake inhibitors in RSBD-like disorder. CONCLUSIONS: Sleepwalking, SRED, RSBD-like disorder, and SRI were common and underrecognized among the psychiatric population in this study. Their occurrences were likely contributed by interacting effect of mental illnesses, sleep disturbances, and specific psychotropic medications. Further prospective study is warranted for clarification of the etiology and clinical management of these potentially dangerous and "hidden" parasomnias.

Status cataplecticus leading to the obstetric complication of prolonged labor.

A 32-year-old married Chinese woman with cataplectic narcolepsy developed status cataplecticus during childbirth. It resulted in prolonged labor and required an emergency cesarean section. The patient had a history of severe and prolonged cataplectic attacks during sexual excitement but relatively mild attacks in other situations. We postulated that her susceptibility to genital stimulation might predispose her to cataplectic attacks during childbirth. Awareness of this potential obstetric complication, especially in those with a history of prominent sex-related cataplexy, together with careful planning and monitoring during the labor process may help to minimize the obstetric risk of patients with narcolepsy.

Longitudinal follow up of primary insomnia patients in a psychiatric clinic.

OBJECTIVE: Insomnia could be a symptom of underlying psychiatric or physical disorder, a risk factor for other psychiatric disorder, or a discrete psychiatric disorder per se. In order to determine the nosological status of primary insomnia, an outcome study was carried out to investigate its diagnostic stability and its relationship to subsequent psychiatric disorders. METHODS: Fifty-three primary insomnia patients in a university hospital psychiatric outpatient clinic were assessed by retrospective case note review, followed by a 6 month prospective follow up with Structured Clinical Interview Schedule for DSM-IV In-patient (SCID-I/P, version 2.0, and a sleep questionnaire. RESULTS: The majority of patients (n =44, 83%) did not develop other psychiatric disorders after 13.4+/-1.2 years from the onset of insomnia. Nine patients (17%) developed mood disorder (n =6), anxiety disorder (n =2) and somatoform disorder (n =1) at 6.3+/-2.3 years after the onset of insomnia. Subjective deterioration of insomnia and a shorter duration of sleep symptoms at the first consultation were associated with the development of secondary psychiatric disorders. Approximately one-third (n =17, 32%) reported symptoms improvement and six (13.2%) were free from medications. Better education was the only factor that predicted improvement in symptoms. CONCLUSIONS: There existed a longitudinal diagnostic stability of primary insomnia in a majority of clinical patients. However, in a proportion of patients, it might either be a risk factor or a prodrome of mood or anxiety disorders.

Fahr's disease: a differential diagnosis of frontal lobe syndrome.

Fahr's disease refers to a rare syndrome characterised by symmetrical and bilateral intracranial calcification. The basal ganglia are the most common site of involvement and most cases present with extra-pyramidal symptoms. We describe two men with Fahr's diseases who presented with prominent frontal lobe symptoms. The first man presented with frequent uncontrollable bursts of laughter and crying spells. He later developed mild dysarthric speech and choreoathetoid movement. The second man presented with progressive changes in personality and behaviour. In both cases, there were no parkinsonian features. Computed tomographic scans of both patients demonstrated extensive symmetrical calcification over the basal ganglia and dentate nuclei. A repeated imaging scan in the second patient revealed progressive cerebral atrophy but reduction in the calcification. No underlying cause for the bilateral calcification was found. As frontal lobe symptoms are usually inconspicuous in the early stage, the presence of these symptoms might be overlooked in clinical practice when compared with those suffering from prominent movement disorders.

The prevalence of narcolepsy among Chinese in Hong Kong.

Narcolepsy is a lifelong, crippling sleep disorder. Although the discovery of the hypocretin system has been a breakthough in genetics, the epidemiological aspects of narcolepsy remain elusive. Ethnic predisposition was suggested to partially account for the 2,500-fold difference in the reported prevalence rates of narcolepsy between Japanese (0.59%) and Israeli Jews (0.00023%). We carried out a general population study, conducting a random telephone survey with a structured questionnaire, which included a validated screening instrument (a Chinese version of the Ullanlinna Narcolepsy Scale). It was followed by clinical-polysomnographic-HLA confirmation of the subjects determined to be positive for narcolepsy based on the questionnaire. Of 9,851 subjects interviewed, 28 subjects (0.28%, 58% female) were screened positive. Ninety percent had a second detailed interview, 64% had HLA typing, and over half of them had a sleep assessment. Only three subjects were found to have genuine narcolepsy. The most common nonnarcolepsy diagnoses were sleep apnea syndrome and sleep-wake schedule disorder. The prevalence rate of narcolepsy in Southern (Hong Kong) Chinese was found to be 0.034% (95% confidence interval = 0.010-0.117%). All available narcoleptic subjects were HLA DRB1-1501 positive and 50% were DQB1-0602 positive. The prevalence rate of narcolepsy among Chinese is comparable to the rates for other populations in studies with stringent epidemiological designs, suggesting that major cross-ethnic differences in the prevalence rates of narcolepsy previously reported likely resulted from methodological limitations.