Virtual Trials: Causally-validated treatment effects efficiently learned from an observational cancer registry.

Randomized controlled trials (RCTs) offer a clear causal interpretation of treatment effects, but are inefficient in terms of information gain per patient. Moreover, because they are intended to test cohort-level effects, RCTs rarely provide information to support precision medicine, which strives to choose the best treatment for an individual patient. If causal information could be efficiently extracted from widely available real-world data, the rapidity of treatment validation could be increased, and its costs reduced. Moreover, inferences could be made across larger, more diverse patient populations. We created a "virtual trial" by fitting a multilevel Bayesian survival model to treatment and outcome records self-reported by 451 brain cancer patients. The model recovers group-level treatment effects comparable to RCTs representing over 3200 patients. The model additionally discovers the feature-treatment interactions needed to make individual-level predictions for precision medicine. By learning from heterogeneous real-world data, virtual trials can generate more causal estimates with fewer patients than RCTs, and they can do so without artificially limiting the patient population. This demonstrates the value of virtual trials as a complement to large randomized controlled trials, especially in highly heterogeneous or rare diseases.

Is Cancer Solvable? Towards Efficient and Ethical Biomedical Science.

Global Cumulative Treatment Analysis (GCTA) is a novel clinical research model combining expert knowledge, and treatment coordination based upon global information-gain, to treat every patient optimally while efficiently searching the vast space that is the realm of cancer research.

Prototyping a precision oncology 3.0 rapid learning platform.

BACKGROUND: We describe a prototype implementation of a platform that could underlie a Precision Oncology Rapid Learning system. RESULTS: We describe the prototype platform, and examine some important issues and details. In the Appendix we provide a complete walk-through of the prototype platform. CONCLUSIONS: The design choices made in this implementation rest upon ten constitutive hypotheses, which, taken together, define a particular view of how a rapid learning medical platform might be defined, organized, and implemented.

Rapid learning for precision oncology.

The emerging paradigm of Precision Oncology 3.0 uses panomics and sophisticated methods of statistical reverse engineering to hypothesize the putative networks that drive a given patient's tumour, and to attack these drivers with combinations of targeted therapies. Here, we review a paradigm termed Rapid Learning Precision Oncology wherein every treatment event is considered as a probe that simultaneously treats the patient and provides an opportunity to validate and refine the models on which the treatment decisions are based. Implementation of Rapid Learning Precision Oncology requires overcoming a host of challenges that include developing analytical tools, capturing the information from each patient encounter and rapidly extrapolating it to other patients, coordinating many patient encounters to efficiently search for effective treatments, and overcoming economic, social and structural impediments, such as obtaining access to, and reimbursement for, investigational drugs.

Groups: knowledge spreadsheets for symbolic biocomputing.

Knowledge spreadsheets (KSs) are a visual tool for interactive data analysis and exploration. They differ from traditional spreadsheets in that rather than being oriented toward numeric data, they work with symbolic knowledge representation structures and provide operations that take into account the semantics of the application domain. 'Groups' is an implementation of KSs within the Pathway Tools system. Groups allows Pathway Tools users to define a group of objects (e.g. groups of genes or metabolites) from a Pathway/Genome Database. Groups can be transformed (e.g. by transforming a metabolite group to the group of pathways in which those metabolites are substrates); combined through set operations; analysed (e.g. through enrichment analysis); and visualized (e.g. by painting onto a metabolic map diagram). Users of the Pathway Tools-based BioCyc.org website have made extensive use of Groups, and an informal survey of Groups users suggests that Groups has achieved the goal of allowing biologists themselves to perform some data manipulations that previously would have required the assistance of a programmer. Database URL: BioCyc.org.

A novel classification of lung cancer into molecular subtypes.

The remarkably heterogeneous nature of lung cancer has become more apparent over the last decade. In general, advanced lung cancer is an aggressive malignancy with a poor prognosis. The discovery of multiple molecular mechanisms underlying the development, progression, and prognosis of lung cancer, however, has created new opportunities for targeted therapy and improved outcome. In this paper, we define "molecular subtypes" of lung cancer based on specific actionable genetic aberrations. Each subtype is associated with molecular tests that define the subtype and drugs that may potentially treat it. We hope this paper will be a useful guide to clinicians and researchers alike by assisting in therapy decision making and acting as a platform for further study. In this new era of cancer treatment, the 'one-size-fits-all' paradigm is being forcibly pushed aside-allowing for more effective, personalized oncologic care to emerge.

A melanoma molecular disease model.

While advanced melanoma remains one of the most challenging cancers, recent developments in our understanding of the molecular drivers of this disease have uncovered exciting opportunities to guide personalized therapeutic decisions. Genetic analyses of melanoma have uncovered several key molecular pathways that are involved in disease onset and progression, as well as prognosis. These advances now make it possible to create a "Molecular Disease Model" (MDM) for melanoma that classifies individual tumors into molecular subtypes (in contrast to traditional histological subtypes), with proposed treatment guidelines for each subtype including specific assays, drugs, and clinical trials. This paper describes such a Melanoma Molecular Disease Model reflecting the latest scientific, clinical, and technological advances.

Targeted Therapy Database (TTD): a model to match patient's molecular profile with current knowledge on cancer biology.

BACKGROUND: The efficacy of current anticancer treatments is far from satisfactory and many patients still die of their disease. A general agreement exists on the urgency of developing molecularly targeted therapies, although their implementation in the clinical setting is in its infancy. In fact, despite the wealth of preclinical studies addressing these issues, the difficulty of testing each targeted therapy hypothesis in the clinical arena represents an intrinsic obstacle. As a consequence, we are witnessing a paradoxical situation where most hypotheses about the molecular and cellular biology of cancer remain clinically untested and therefore do not translate into a therapeutic benefit for patients. OBJECTIVE: To present a computational method aimed to comprehensively exploit the scientific knowledge in order to foster the development of personalized cancer treatment by matching the patient's molecular profile with the available evidence on targeted therapy. METHODS: To this aim we focused on melanoma, an increasingly diagnosed malignancy for which the need for novel therapeutic approaches is paradigmatic since no effective treatment is available in the advanced setting. Relevant data were manually extracted from peer-reviewed full-text original articles describing any type of anti-melanoma targeted therapy tested in any type of experimental or clinical model. To this purpose, Medline, Embase, Cancerlit and the Cochrane databases were searched. RESULTS AND CONCLUSIONS: We created a manually annotated database (Targeted Therapy Database, TTD) where the relevant data are gathered in a formal representation that can be computationally analyzed. Dedicated algorithms were set up for the identification of the prevalent therapeutic hypotheses based on the available evidence and for ranking treatments based on the molecular profile of individual patients. In this essay we describe the principles and computational algorithms of an original method developed to fully exploit the available knowledge on cancer biology with the ultimate goal of fruitfully driving both preclinical and clinical research on anticancer targeted therapy. In the light of its theoretical nature, the prediction performance of this model must be validated before it can be implemented in the clinical setting.

The promise and perils of pre-publication review: a multi-agent simulation of biomedical discovery under varying levels of review stringency.

The Internet has enabled profound changes in the way science is performed, especially in scientific communications. Among the most important of these changes is the possibility of new models for pre-publication review, ranging from the current, relatively strict peer-review model, to entirely unreviewed, instant self-publication. Different models may affect scientific progress by altering both the quality and quantity of papers available to the research community. To test how models affect the community, I used a multi-agent simulation of treatment selection and outcome in a patient population to examine how various levels of pre-publication review might affect the rate of scientific progress. I identified a "sweet spot" between the points of very limited and very strict requirements for pre-publication review. The model also produced a u-shaped curve where very limited review requirement was slightly superior to a moderate level of requirement, but not as large as the aforementioned sweet spot. This unexpected phenomenon appears to result from the community taking longer to discover the correct treatment with more strict pre-publication review. In the parameter regimens I explored, both completely unreviewed and very strictly reviewed scientific communication seems likely to hinder scientific progress. Much more investigation is warranted. Multi-agent simulations can help to shed light on complex questions of scientific communication and exhibit interesting, unexpected behaviors.

Soccer science and the Bayes community: exploring the cognitive implications of modern scientific communication.

Science is a form of distributed analysis involving both individual work that produces new knowledge and collaborative work to exchange information with the larger community. There are many particular ways in which individual and community can interact in science, and it is difficult to assess how efficient these are, and what the best way might be to support them. This paper reports on a series of experiments in this area and a prototype implementation using a research platform called CACHE. CACHE both supports experimentation with different structures of interaction between individual and community cognition and serves as a prototype for computational support for those structures. We particularly focus on CACHE-BC, the Bayes community version of CACHE, within which the community can break up analytical tasks into "mind-sized" units and use provenance tracking to keep track of the relationship between these units.

BioBIKE: a Web-based, programmable, integrated biological knowledge base.

BioBIKE (biobike.csbc.vcu.edu) is a web-based environment enabling biologists with little programming expertise to combine tools, data, and knowledge in novel and possibly complex ways, as demanded by the biological problem at hand. BioBIKE is composed of three integrated components: a biological knowledge base, a graphical programming interface and an extensible set of tools. Each of the five current BioBIKE instances provides all available information (genomic, metabolic, experimental) appropriate to a given research community. The BioBIKE programming language and graphical programming interface employ familiar operations to help users combine functions and information to conduct biologically meaningful analyses. Many commonly used tools, such as Blast and PHYLIP, are built-in, allowing users to access them within the same interface and to pass results from one to another. Users may also invent their own tools, packaging complex expressions under a single name, which is immediately made accessible through the graphical interface. BioBIKE represents a partial solution to the difficult question of how to enable those with no background in computer programming to work directly and creatively with mass biological information. BioBIKE is distributed under the MIT Open Source license. A description of the underlying language and other technical matters is available at www.Biobike.org.

Extraction of Conditional Probabilities of the Relationships Between Drugs, Diseases, and Genes from PubMed Guided by Relationships in PharmGKB.

Guided by curated associations between genes, treatments (i.e., drugs), and diseases in pharmGKB, we constructed n-way Bayesian networks based on conditional probability tables (cpt's) extracted from co-occurrence statistics over the entire Pubmed corpus, producing a broad-coverage analysis of the relationships between these biological entities. The networks suggest hypotheses regarding drug mechanisms, treatment biomarkers, and/or potential markers of genetic disease. The cpt's enable Trio, an inferential database, to query indirect (inferred) relationships via an SQL-like query language.

Alternative photosynthetic electron flow to oxygen in marine Synechococcus.

Cyanobacteria dominate the world's oceans where iron is often barely detectable. One manifestation of low iron adaptation in the oligotrophic marine environment is a decrease in levels of iron-rich photosynthetic components, including the reaction center of photosystem I and the cytochrome b6f complex [R.F. Strzepek and P.J. Harrison, Photosynthetic architecture differs in coastal and oceanic diatoms, Nature 431 (2004) 689-692.]. These thylakoid membrane components have well characterised roles in linear and cyclic photosynthetic electron transport and their low abundance creates potential impediments to photosynthetic function. Here we show that the marine cyanobacterium Synechococcus WH8102 exhibits significant alternative electron flow to O2, a potential adaptation to the low iron environment in oligotrophic oceans. This alternative electron flow appears to extract electrons from the intersystem electron transport chain, prior to photosystem I. Inhibitor studies demonstrate that a propyl gallate-sensitive oxidase mediates this flow of electrons to oxygen, which in turn alleviates excessive photosystem II excitation pressure that can often occur even at relatively low irradiance. These findings are also discussed in the context of satisfying the energetic requirements of the cell when photosystem I abundance is low.

UNDERSTANDING NITROGEN LIMITATION IN AUREOCOCCUS ANOPHAGEFFERENS (PELAGOPHYCEAE) THROUGH cDNA AND qRT-PCR ANALYSIS(1).

Brown tides of the marine pelagophyte Aureococcus anophagefferens Hargraves et Sieburth have been investigated extensively for the past two decades. Its growth is fueled by a variety of nitrogen (N) compounds, with dissolved organic nitrogen (DON) being particularly important during blooms. Characterization of a cDNA library suggests that A. anophagefferens can assimilate eight different forms of N. Expression of genes related to the sensing, uptake, and assimilation of inorganic and organic N, as well as the catabolic process of autophagy, was assayed in cells grown on different N sources and in N-limited cells. Growth on nitrate elicited an increase in the relative expression of nitrate and ammonium transporters, a nutrient stress-induced transporter, and a sensory kinase. Growth on urea increased the relative expression of a urea and a formate/nitrite transporter, while growth on ammonium resulted in an increase in the relative expression of an ammonium transporter, a novel ATP-binding cassette (ABC) transporter and a putative high-affinity phosphate transporter. N limitation resulted in a 30- to 110-fold increase in the relative expression of nitrate, ammonium, urea, amino acid/polyamine, and formate/nitrite transporters. A. anophagefferens demonstrated the highest relative accumulation of a transcript encoding a novel purine transporter, which was highly expressed across all N sources. This finding suggests that purines are an important source of N for the growth of this organism and could possibly contribute to the initiation and maintenance of blooms in the natural environment.

Deductive biocomputing.

BACKGROUND: As biologists increasingly rely upon computational tools, it is imperative that they be able to appropriately apply these tools and clearly understand the methods the tools employ. Such tools must have access to all the relevant data and knowledge and, in some sense, "understand" biology so that they can serve biologists' goals appropriately and "explain" in biological terms how results are computed. METHODOLOGY/PRINCIPAL FINDINGS: We describe a deduction-based approach to biocomputation that semiautomatically combines knowledge, software, and data to satisfy goals expressed in a high-level biological language. The approach is implemented in an open source web-based biocomputing platform called BioDeducta, which combines SRI's SNARK theorem prover with the BioBike interactive integrated knowledge base. The biologist/user expresses a high-level conjecture, representing a biocomputational goal query, without indicating how this goal is to be achieved. A subject domain theory, represented in SNARK's logical language, transforms the terms in the conjecture into capabilities of the available resources and the background knowledge necessary to link them together. If the subject domain theory enables SNARK to prove the conjecture--that is, to find paths between the goal and BioBike resources--then the resulting proofs represent solutions to the conjecture/query. Such proofs provide provenance for each result, indicating in detail how they were computed. We demonstrate BioDeducta by showing how it can approximately replicate a previously published analysis of genes involved in the adaptation of cyanobacteria to different light niches. CONCLUSIONS/SIGNIFICANCE: Through the use of automated deduction guided by a biological subject domain theory, this work is a step towards enabling biologists to conveniently and efficiently marshal integrated knowledge, data, and computational tools toward resolving complex biological queries.

EST assembly supported by a draft genome sequence: an analysis of the Chlamydomonas reinhardtii transcriptome.

Clustering and assembly of expressed sequence tags (ESTs) constitute the basis for most genomewide descriptions of a transcriptome. This approach is limited by the decline in sequence quality toward the end of each EST, impacting both sequence clustering and assembly. Here, we exploit the available draft genome sequence of the unicellular green alga Chlamydomonas reinhardtii to guide clustering and to correct errors in the ESTs. We have grouped all available EST and cDNA sequences into 12,063 ACEGs (assembly of contiguous ESTs based on genome) and generated 15,857 contigs of average length 934 nt. We predict that roughly 3000 of our contigs represent full-length transcripts. Compared to previous assemblies, ACEGs show extended contig length, increased accuracy and a reduction in redundancy. Because our assembly protocol also uses ESTs with no corresponding genomic sequences, it provides sequence information for genes interrupted by sequence gaps. Detailed analysis of randomly sampled ACEGs reveals several hundred putative cases of alternative splicing, many overlapping transcription units and new genes not identified by gene prediction algorithms. Our protocol, although developed for and tailored to the C. reinhardtii dataset, can be exploited by any eukaryotic genome project for which both a draft genome sequence and ESTs are available.

Constructing explanatory process models from biological data and knowledge.

OBJECTIVE: We address the task of inducing explanatory models from observations and knowledge about candidate biological processes, using the illustrative problem of modeling photosynthesis regulation. METHODS: We cast both models and background knowledge in terms of processes that interact to account for behavior. We also describe IPM, an algorithm for inducing quantitative process models from such input. RESULTS: We demonstrate IPM's use both on photosynthesis and on a second domain, biochemical kinetics, reporting the models induced and their fit to observations. CONCLUSION: We consider the generality of our approach, discuss related research on biological modeling, and suggest directions for future work.

Generation of an oligonucleotide array for analysis of gene expression in Chlamydomonas reinhardtii.

The availability of genome sequences makes it possible to develop microarrays that can be used for profiling gene expression over developmental time, as organisms respond to environmental challenges, and for comparison between wild-type and mutant strains under various conditions. The desired characteristics of microarrays (intense signals, hybridization specificity and extensive coverage of the transcriptome) were not fully met by the previous Chlamydomonas reinhardtii microarray: probes derived from cDNA sequences (approximately 300 bp) were prone to some nonspecific cross-hybridization and coverage of the transcriptome was only approximately 20%. The near completion of the C. reinhardtii nuclear genome sequence and the availability of extensive cDNA information have made it feasible to improve upon these aspects. After developing a protocol for selecting a high-quality unigene set representing all known expressed sequences, oligonucleotides were designed and a microarray with approximately 10,000 unique array elements (approximately 70 bp) covering 87% of the known transcriptome was developed. This microarray will enable researchers to generate a global view of gene expression in C. reinhardtii. Furthermore, the detailed description of the protocol for selecting a unigene set and the design of oligonucleotides may be of interest for laboratories interested in developing microarrays for organisms whose genome sequences are not yet completed (but are nearing completion).

BioLingua: a programmable knowledge environment for biologists.

UNLABELLED: BioLingua is an interactive, web-based programming environment that enables biologists to analyze biological systems by combining knowledge and data through direct end-user programming. BioLingua embeds a mature symbolic programming language in a frame-based knowledge environment, integrating genomic and pathway knowledge about a class of similar organisms. The BioLingua language provides interfaces to numerous state-of-the-art bioinformatic tools, making these available as an integrated package through the novel use of web-based programmability and an integrated Wiki-based community code and data store. The pilot instantiation of BioLingua, which has been developed in collaboration with several cyanobacteriologists, integrates knowledge about a subset of cyanobacteria with the Gene Ontology, KEGG and BioCyc knowledge bases. We introduce the BioLingua concept, architecture and language, and give several examples of its use in complex analyses. AVAILABILITY: Extensive documentation is available online at http://nostoc.stanford.edu/Docs/index.html CONTACT: JShrager@Stanford.edu