Pharos: Collating protein information to shed light on the druggable genome.

The 'druggable genome' encompasses several protein families, but only a subset of targets within them have attracted significant research attention and thus have information about them publicly available. The Illuminating the Druggable Genome (IDG) program was initiated in 2014, has the goal of developing experimental techniques and a Knowledge Management Center (KMC) that would collect and organize information about protein targets from four families, representing the most common druggable targets with an emphasis on understudied proteins. Here, we describe two resources developed by the KMC: the Target Central Resource Database (TCRD) which collates many heterogeneous gene/protein datasets and Pharos (https://pharos.nih.gov), a multimodal web interface that presents the data from TCRD. We briefly describe the types and sources of data considered by the KMC and then highlight features of the Pharos interface designed to enable intuitive access to the IDG knowledgebase. The aim of Pharos is to encourage 'serendipitous browsing', whereby related, relevant information is made easily discoverable. We conclude by describing two use cases that highlight the utility of Pharos and TCRD.

TIN-X: target importance and novelty explorer.

MOTIVATION: The increasing amount of peer-reviewed manuscripts requires the development of specific mining tools to facilitate the visual exploration of evidence linking diseases and proteins. RESULTS: We developed TIN-X, the Target Importance and Novelty eXplorer, to visualize the association between proteins and diseases, based on text mining data processed from scientific literature. In the current implementation, TIN-X supports exploration of data for G-protein coupled receptors, kinases, ion channels, and nuclear receptors. TIN-X supports browsing and navigating across proteins and diseases based on ontology classes, and displays a scatter plot with two proposed new bibliometric statistics: Importance and Novelty. AVAILABILITY AND IMPLEMENTATION: http://www.newdrugtargets.org. CONTACT: cbologa@salud.unm.edu.

Protein biomarker druggability profiling.

Developing automated and interactive methods for building a model by incorporating mechanistic and potentially causal annotations of ranked biomarkers of a disease or clinical condition followed by a mapping into a contextual framework in disease-linked biochemical pathways can be used for potential drug-target evaluation and for proposing new drug targets. We demonstrate the potential of this approach using ranked protein biomarkers obtained in neonatal sepsis by enrolling 127 infants (39 infants with late onset neonatal sepsis and 88 control infants) and by performing a focused proteomic profile of the sera and by applying the interactive druggability profiling algorithm (DPA) developed by us.

Applying active learning to assertion classification of concepts in clinical text.

Supervised machine learning methods for clinical natural language processing (NLP) research require a large number of annotated samples, which are very expensive to build because of the involvement of physicians. Active learning, an approach that actively samples from a large pool, provides an alternative solution. Its major goal in classification is to reduce the annotation effort while maintaining the quality of the predictive model. However, few studies have investigated its uses in clinical NLP. This paper reports an application of active learning to a clinical text classification task: to determine the assertion status of clinical concepts. The annotated corpus for the assertion classification task in the 2010 i2b2/VA Clinical NLP Challenge was used in this study. We implemented several existing and newly developed active learning algorithms and assessed their uses. The outcome is reported in the global ALC score, based on the Area under the average Learning Curve of the AUC (Area Under the Curve) score. Results showed that when the same number of annotated samples was used, active learning strategies could generate better classification models (best ALC-0.7715) than the passive learning method (random sampling) (ALC-0.7411). Moreover, to achieve the same classification performance, active learning strategies required fewer samples than the random sampling method. For example, to achieve an AUC of 0.79, the random sampling method used 32 samples, while our best active learning algorithm required only 12 samples, a reduction of 62.5% in manual annotation effort.

Unexplored therapeutic opportunities in the human genome.

A large proportion of biomedical research and the development of therapeutics is focused on a small fraction of the human genome. In a strategic effort to map the knowledge gaps around proteins encoded by the human genome and to promote the exploration of currently understudied, but potentially druggable, proteins, the US National Institutes of Health launched the Illuminating the Druggable Genome (IDG) initiative in 2014. In this article, we discuss how the systematic collection and processing of a wide array of genomic, proteomic, chemical and disease-related resource data by the IDG Knowledge Management Center have enabled the development of evidence-based criteria for tracking the target development level (TDL) of human proteins, which indicates a substantial knowledge deficit for approximately one out of three proteins in the human proteome. We then present spotlights on the TDL categories as well as key drug target classes, including G protein-coupled receptors, protein kinases and ion channels, which illustrate the nature of the unexplored opportunities for biomedical research and therapeutic development.

Unexplored therapeutic opportunities in the human genome.

This corrects the article DOI: 10.1038/nrd.2018.14.

Learning causal and predictive clinical practice guidelines from data.

Clinical practice guidelines (CPG) propose preventive, diagnostic and treatment strategies based on the best available evidence. CPG enable practice of evidencebased medicine and bring about standardization of healthcare delivery in a given hospital, region, country or the whole world. This study explores generation of guidelines from data using machine learning, causal discovery methods and the domain of high blood pressure as an example.

Formalizing drug indications on the road to therapeutic intent.

Therapeutic intent, the reason behind the choice of a therapy and the context in which a given approach should be used, is an important aspect of medical practice. There are unmet needs with respect to current electronic mapping of drug indications. For example, the active ingredient sildenafil has 2 distinct indications, which differ solely on dosage strength. In progressing toward a practice of precision medicine, there is a need to capture and structure therapeutic intent for computational reuse, thus enabling more sophisticated decision-support tools and a possible mechanism for computer-aided drug repurposing. The indications for drugs, such as those expressed in the Structured Product Labels approved by the US Food and Drug Administration, appears to be a tractable area for developing an application ontology of therapeutic intent.

Causal discovery using a Bayesian local causal discovery algorithm.

This study focused on the development and application of an efficient algorithm to induce causal relationships from observational data. The algorithm, called BLCD, is based on a causal Bayesian network framework. BLCD initially uses heuristic greedy search to derive the Markov Blanket (MB) of a node that serves as the "locality" for the identification of pair-wise causal relationships. BLCD takes as input a dataset and outputs potential causes of the form variable X causally influences variable Y. Identification of the causal factors of diseases and outcomes, can help formulate better management, prevention and control strategies for the improvement of health care. In this study we focused on investigating factors that may contribute causally to infant mortality in the United States. We used the U.S. Linked Birth/Infant Death dataset for 1991 with more than four million records and about 200 variables for each record. Our sample consisted of 41,155 re-cords randomly selected from the whole dataset. Each record had maternal, paternal and child factors and the outcome at the end of the first year--whether the infant survived or not. Using the infant birth and death dataset as input, BLCD out-put six purported causal relationships. Three out of the six relationships seem plausible. Even though we have not yet discovered a clinically novel causal link, we plan to look for novel causal pathways using the full sample.

Medical decision support using machine learning for early detection of late-onset neonatal sepsis.

OBJECTIVE: The objective was to develop non-invasive predictive models for late-onset neonatal sepsis from off-the-shelf medical data and electronic medical records (EMR). DESIGN: The data used in this study are from 299 infants admitted to the neonatal intensive care unit in the Monroe Carell Jr. Children's Hospital at Vanderbilt and evaluated for late-onset sepsis. Gold standard diagnostic labels (sepsis negative, culture positive sepsis, culture negative/clinical sepsis) were assigned based on all the laboratory, clinical and microbiology data available in EMR. Only data that were available up to 12 h after phlebotomy for blood culture testing were used to build predictive models using machine learning (ML) algorithms. MEASUREMENT: We compared sensitivity, specificity, positive predictive value and negative predictive value of sepsis treatment of physicians with the predictions of models generated by ML algorithms. RESULTS: The treatment sensitivity of all the nine ML algorithms and specificity of eight out of the nine ML algorithms tested exceeded that of the physician when culture-negative sepsis was included. When culture-negative sepsis was excluded both sensitivity and specificity exceeded that of the physician for all the ML algorithms. The top three predictive variables were the hematocrit or packed cell volume, chorioamnionitis and respiratory rate. CONCLUSIONS: Predictive models developed from off-the-shelf and EMR data using ML algorithms exceeded the treatment sensitivity and treatment specificity of clinicians. A prospective study is warranted to assess the clinical utility of the ML algorithms in improving the accuracy of antibiotic use in the management of neonatal sepsis.

Machine learning for predicting the response of breast cancer to neoadjuvant chemotherapy.

OBJECTIVE: To employ machine learning methods to predict the eventual therapeutic response of breast cancer patients after a single cycle of neoadjuvant chemotherapy (NAC). MATERIALS AND METHODS: Quantitative dynamic contrast-enhanced MRI and diffusion-weighted MRI data were acquired on 28 patients before and after one cycle of NAC. A total of 118 semiquantitative and quantitative parameters were derived from these data and combined with 11 clinical variables. We used Bayesian logistic regression in combination with feature selection using a machine learning framework for predictive model building. RESULTS: The best predictive models using feature selection obtained an area under the curve of 0.86 and an accuracy of 0.86, with a sensitivity of 0.88 and a specificity of 0.82. DISCUSSION: With the numerous options for NAC available, development of a method to predict response early in the course of therapy is needed. Unfortunately, by the time most patients are found not to be responding, their disease may no longer be surgically resectable, and this situation could be avoided by the development of techniques to assess response earlier in the treatment regimen. The method outlined here is one possible solution to this important clinical problem. CONCLUSIONS: Predictive modeling approaches based on machine learning using readily available clinical and quantitative MRI data show promise in distinguishing breast cancer responders from non-responders after the first cycle of NAC.

Type 2 diabetes risk forecasting from EMR data using machine learning.

OBJECTIVE: To test the feasibility of using data collected in electronic medical records for development of effective models for diabetes risk forecasting. METHODS: Using available demographic, clinical and lab parameters of more than two thousand patients from Electronic medical records, we applied different machine learning algorithms to assess the risk of development of type 2 diabetes (T2D) six months to one year later. RESULTS: We achieved an AUC greater than 0.8 for predicting type 2 diabetes 365 days and 180 days prior to diagnosis of diabetes. CONCLUSION: Diabetes risk forecasting using data from EMR is innovative and has the potential to identify, automatically, high-risk populations for early intervention with life style modifications such as diet and exercise to prevent or delay the development of T2D. Our study shows that T2D risk forecasting from EMR data is feasible.

A study of machine-learning-based approaches to extract clinical entities and their assertions from discharge summaries.

OBJECTIVE: The authors' goal was to develop and evaluate machine-learning-based approaches to extracting clinical entities-including medical problems, tests, and treatments, as well as their asserted status-from hospital discharge summaries written using natural language. This project was part of the 2010 Center of Informatics for Integrating Biology and the Bedside/Veterans Affairs (VA) natural-language-processing challenge. DESIGN: The authors implemented a machine-learning-based named entity recognition system for clinical text and systematically evaluated the contributions of different types of features and ML algorithms, using a training corpus of 349 annotated notes. Based on the results from training data, the authors developed a novel hybrid clinical entity extraction system, which integrated heuristic rule-based modules with the ML-base named entity recognition module. The authors applied the hybrid system to the concept extraction and assertion classification tasks in the challenge and evaluated its performance using a test data set with 477 annotated notes. MEASUREMENTS: Standard measures including precision, recall, and F-measure were calculated using the evaluation script provided by the Center of Informatics for Integrating Biology and the Bedside/VA challenge organizers. The overall performance for all three types of clinical entities and all six types of assertions across 477 annotated notes were considered as the primary metric in the challenge. RESULTS AND DISCUSSION: Systematic evaluation on the training set showed that Conditional Random Fields outperformed Support Vector Machines, and semantic information from existing natural-language-processing systems largely improved performance, although contributions from different types of features varied. The authors' hybrid entity extraction system achieved a maximum overall F-score of 0.8391 for concept extraction (ranked second) and 0.9313 for assertion classification (ranked fourth, but not statistically different than the first three systems) on the test data set in the challenge.

Detecting abbreviations in discharge summaries using machine learning methods.

Recognition and identification of abbreviations is an important, challenging task in clinical natural language processing (NLP). A comprehensive lexical resource comprised of all common, useful clinical abbreviations would have great applicability. The authors present a corpus-based method to create a lexical resource of clinical abbreviations using machine-learning (ML) methods, and tested its ability to automatically detect abbreviations from hospital discharge summaries. Domain experts manually annotated abbreviations in seventy discharge summaries, which were randomly broken into a training set (40 documents) and a test set (30 documents). We implemented and evaluated several ML algorithms using the training set and a list of pre-defined features. The subsequent evaluation using the test set showed that the Random Forest classifier had the highest F-measure of 94.8% (precision 98.8% and recall of 91.2%). When a voting scheme was used to combine output from various ML classifiers, the system achieved the highest F-measure of 95.7%.

Early prediction of the response of breast tumors to neoadjuvant chemotherapy using quantitative MRI and machine learning.

The ability to predict early in the course of treatment the response of breast tumors to neoadjuvant chemotherapy can stratify patients based on response for patient-specific treatment strategies. Currently response to neoadjuvant chemotherapy is evaluated based on physical exam or breast imaging (mammogram, ultrasound or conventional breast MRI). There is a poor correlation among these measurements and with the actual tumor size when measured by the pathologist during definitive surgery. We tested the feasibility of using quantitative MRI as a tool for early prediction of tumor response. Between 2007 and 2010 twenty consecutive patients diagnosed with Stage II/III breast cancer and receiving neoadjuvant chemotherapy were enrolled on a prospective imaging study. Our study showed that quantitative MRI parameters along with routine clinical measures can predict responders from non-responders to neoadjuvant chemotherapy. The best predictive model had an accuracy of 0.9, a positive predictive value of 0.91 and an AUC of 0.96.

Extracting and integrating data from entire electronic health records for detecting colorectal cancer cases.

Identification of a cohort of patients with specific diseases is an important step for clinical research that is based on electronic health records (EHRs). Informatics approaches combining structured EHR data, such as billing records, with narrative text data have demonstrated utility for such tasks. This paper describes an algorithm combining machine learning and natural language processing to detect patients with colorectal cancer (CRC) from entire EHRs at Vanderbilt University Hospital. We developed a general case detection method that consists of two steps: 1) extraction of positive CRC concepts from all clinical notes (document-level concept identification); and 2) determination of CRC cases using aggregated information from both clinical narratives and structured billing data (patient-level case determination). For each step, we compared performance of rule-based and machine-learning-based approaches. Using a manually reviewed data set containing 300 possible CRC patients (150 for training and 150 for testing), we showed that our method achieved F-measures of 0.996 for document level concept identification, and 0.93 for patient level case detection.

Early prediction of reading disability using machine learning.

This paper presents application of machine learning methods on a 356 sample dataset for early prediction of reading disability among first graders. A wide array of classifiers consisting of Support Vector Machines, Decision Trees (CART and C4.5), Linear Discriminant Analysis, k Nearest Neighbor and Naïve Bayes Classifiers were used in this study. Markov Blanket based feature selection algorithms (HITON-PC and HITON-MB) and wrapper based feature selection algorithms (forward, backward, forward and backward wrapping algorithm and support vector machine recursive feature elimination) were used to select the most relevant features for classification. The results indicate that an AUC score greater than 0.9 can be achieved using SVM classifiers even with a small set of demographics and screening variables. Moreover, a method for generating expert interpretable decision tree models from the high accuracy SVM models is also presented.