Serial KinderMiner (SKiM) discovers and annotates biomedical knowledge using co-occurrence and transformer models.

BACKGROUND: The PubMed archive contains more than 34 million articles; consequently, it is becoming increasingly difficult for a biomedical researcher to keep up-to-date with different knowledge domains. Computationally efficient and interpretable tools are needed to help researchers find and understand associations between biomedical concepts. The goal of literature-based discovery (LBD) is to connect concepts in isolated literature domains that would normally go undiscovered. This usually takes the form of an A-B-C relationship, where A and C terms are linked through a B term intermediate. Here we describe Serial KinderMiner (SKiM), an LBD algorithm for finding statistically significant links between an A term and one or more C terms through some B term intermediate(s). The development of SKiM is motivated by the observation that there are only a few LBD tools that provide a functional web interface, and that the available tools are limited in one or more of the following ways: (1) they identify a relationship but not the type of relationship, (2) they do not allow the user to provide their own lists of B or C terms, hindering flexibility, (3) they do not allow for querying thousands of C terms (which is crucial if, for instance, the user wants to query connections between a disease and the thousands of available drugs), or (4) they are specific for a particular biomedical domain (such as cancer). We provide an open-source tool and web interface that improves on all of these issues. RESULTS: We demonstrate SKiM's ability to discover useful A-B-C linkages in three control experiments: classic LBD discoveries, drug repurposing, and finding associations related to cancer. Furthermore, we supplement SKiM with a knowledge graph built with transformer machine-learning models to aid in interpreting the relationships between terms found by SKiM. Finally, we provide a simple and intuitive open-source web interface ( https://skim.morgridge.org ) with comprehensive lists of drugs, diseases, phenotypes, and symptoms so that anyone can easily perform SKiM searches. CONCLUSIONS: SKiM is a simple algorithm that can perform LBD searches to discover relationships between arbitrary user-defined concepts. SKiM is generalized for any domain, can perform searches with many thousands of C term concepts, and moves beyond the simple identification of an existence of a relationship; many relationships are given relationship type labels from our knowledge graph.

Serial KinderMiner (SKiM) Discovers and Annotates Biomedical Knowledge Using Co-Occurrence and Transformer Models.

Background: The PubMed database contains more than 34 million articles; consequently, it is becoming increasingly difficult for a biomedical researcher to keep up-to-date with different knowledge domains. Computationally efficient and interpretable tools are needed to help researchers find and understand associations between biomedical concepts. The goal of literature-based discovery (LBD) is to connect concepts in isolated literature domains that would normally go undiscovered. This usually takes the form of an A-B-C relationship, where A and C terms are linked through a B term intermediate. Here we describe Serial KinderMiner (SKiM), an LBD algorithm for finding statistically significant links between an A term and one or more C terms through some B term intermediate(s). The development of SKiM is motivated by the the observation that there are only a few LBD tools that provide a functional web interface, and that the available tools are limited in one or more of the following ways: 1) they identify a relationship but not the type of relationship, 2) they do not allow the user to provide their own lists of B or C terms, hindering flexibility, 3) they do not allow for querying thousands of C terms (which is crucial if, for instance, the user wants to query connections between a disease and the thousands of available drugs), or 4) they are specific for a particular biomedical domain (such as cancer). We provide an open-source tool and web interface that improves on all of these issues. Results: We demonstrate SKiM's ability to discover useful A-B-C linkages in three control experiments: classic LBD discoveries, drug repurposing, and finding associations related to cancer. Furthermore, we supplement SKiM with a knowledge graph built with transformer machine-learning models to aid in interpreting the relationships between terms found by SKiM. Finally, we provide a simple and intuitive open-source web interface ( https://skim.morgridge.org ) with comprehensive lists of drugs, diseases, phenotypes, and symptoms so that anyone can easily perform SKiM searches. Conclusions: SKiM is a simple algorithm that can perform LBD searches to discover relationships between arbitrary user-defined concepts. SKiM is generalized for any domain, can perform searches with many thousands of C term concepts, and moves beyond the simple identification of an existence of a relationship; many relationships are given relationship type labels from our knowledge graph.

Biomedical Literature Mining for Repurposing Laboratory Tests.

Epidemiological studies identifying biological markers of disease state are valuable, but can be time-consuming, expensive, and require extensive intuition and expertise. Furthermore, not all hypothesized markers will be borne out in a study, suggesting that high-quality initial hypotheses are crucial. In this chapter, we describe a high-throughput pipeline to produce a ranked list of high-quality hypothesized biomarkers for diseases. We review an example use of this approach to generate a large number of candidate disease biomarker hypotheses derived from machine learning models, filter and rank them according to their potential novelty using text mining, and corroborate the most promising hypotheses with further statistical modeling. The example use of the pipeline uses a large electronic health record dataset and the PubMed corpus, to find several promising hypothesized laboratory tests with previously undocumented correlations to particular diseases.

Artificial intelligence-assisted phenotype discovery of fragile X syndrome in a population-based sample.

PURPOSE: Fragile X syndrome (FXS), the most prevalent inherited cause of intellectual disability, remains underdiagnosed in the general population. Clinical studies have shown that individuals with FXS have a complex health profile leading to unique clinical needs. However, the full impact of this X-linked disorder on the health of affected individuals is unclear and the prevalence of co-occurring conditions is unknown. METHODS: We mined the longitudinal electronic health records from more than one million individuals to investigate the health characteristics of patients who have been clinically diagnosed with FXS. Additionally, using machine-learning approaches, we created predictive models to identify individuals with FXS in the general population. RESULTS: Our discovery-oriented approach identified the associations of FXS with a wide range of medical conditions including circulatory, endocrine, digestive, and genitourinary, in addition to mental and neurological disorders. We successfully created predictive models to identify cases five years prior to clinical diagnosis of FXS without relying on any genetic or familial data. CONCLUSION: Although FXS is often thought of primarily as a neurological disorder, it is in fact a multisystem syndrome involving many co-occurring conditions, some primary and some secondary, and they are associated with a considerable burden on patients and their families.

KinderMiner Web: a simple web tool for ranking pairwise associations in biomedical applications.

Many important scientific discoveries require lengthy experimental processes of trial and error and could benefit from intelligent prioritization based on deep domain understanding. While exponential growth in the scientific literature makes it difficult to keep current in even a single domain, that same rapid growth in literature also presents an opportunity for automated extraction of knowledge via text mining. We have developed a web application implementation of the KinderMiner algorithm for proposing ranked associations between a list of target terms and a key phrase. Any key phrase and target term list can be used for biomedical inquiry. We built the web application around a text index derived from PubMed. It is the first publicly available implementation of the algorithm, is fast and easy to use, and includes an interactive analysis tool. The KinderMiner web application is a public resource offering scientists a cohesive summary of what is currently known about a particular topic within the literature, and helping them to prioritize experiments around that topic. It performs comparably or better to similar state-of-the-art text mining tools, is more flexible, and can be applied to any biomedical topic of interest. It is also continually improving with quarterly updates to the underlying text index and through response to suggestions from the community. The web application is available at https://www.kinderminer.org.

Automated Extraction and Visualization of Protein-Protein Interaction Networks and Beyond: A Text-Mining Protocol.

Proteins perform their functions by interacting with other proteins. Protein-protein interaction (PPI) is critical for understanding the functions of individual proteins, the mechanisms of biological processes, and the disease mechanisms. High-throughput experiments accumulated a huge number of PPIs in PubMed articles, and their extraction is possible only through automated approaches. The standard text-mining protocol includes four major tasks, namely, recognizing protein mentions, normalizing protein names and aliases to unique identifiers such as gene symbol, extracting PPIs, and visualizing the PPI network using Cytoscape or other visualization tools. Each task is challenging and has been revised over several years to improve the performance. We present a protocol based on our hybrid approaches and show the possibility of presenting each task as an independent web-based tool, NAGGNER for protein name recognition, ProNormz for protein name normalization, PPInterFinder for PPI extraction, and HPIminer for PPI network visualization. The protocol is specific to human but can be generalized to other organisms. We include KinderMiner, our most recent text-mining tool that predicts PPIs by retrieving significant co-occurring protein pairs. The algorithm is simple, easy to implement, and generalizable to other biological challenges.

Data-driven phenotype discovery of FMR1 premutation carriers in a population-based sample.

The impact of the FMR1 premutation on human health is the subject of considerable controversy. A fundamental unanswered question is whether carrying the premutation allele is directly correlated with clinical phenotypes. A challenging problem in past genotype-phenotype studies of the FMR1 premutation is ascertainment bias, which could lead to invalid research conclusions and negatively affect clinical practice. Here, we created the first population-based FMR1-informed biobank to find the pattern of health characteristics in premutation carriers. Our extensive phenotyping shows that premutation carriers experience a clinical profile that is significantly different from controls and is evident throughout adulthood. Comprehensive understanding of the clinical risk associated with this genetic variant is critical for premutation carriers, their families, and clinicians and has important implications for public health.

Machine Learning Assisted Discovery of Novel Predictive Lab Tests Using Electronic Health Record Data.

Epidemiological studies identifying biological markers of disease state are valuable, but can be time-consuming, expensive, and require extensive intuition and expertise. Furthermore, not all hypothesized markers will be borne out in a study, suggesting that higher quality initial hypotheses are crucial. In this work, we propose a high-throughput pipeline to produce a ranked list of high-quality hypothesized marker laboratory tests for diagnoses. Our pipeline generates a large number of candidate lab-diagnosis hypotheses derived from machine learning models, filters and ranks them according to their potential novelty using text mining, and corroborate final hypotheses with logistic regression analysis. We test our approach on a large electronic health record dataset and the PubMed corpus, and find several promising candidate hypotheses.

Machine learning to predict developmental neurotoxicity with high-throughput data from 2D bio-engineered tissues.

There is a growing need for fast and accurate methods for testing developmental neurotoxicity across several chemical exposure sources. Current approaches, such as in vivo animal studies, and assays of animal and human primary cell cultures, suffer from challenges related to time, cost, and applicability to human physiology. Prior work has demonstrated success employing machine learning to predict developmental neurotoxicity using gene expression data collected from human 3D tissue models exposed to various compounds. The 3D model is biologically similar to developing neural structures, but its complexity necessitates extensive expertise and effort to employ. By instead focusing solely on constructing an assay of developmental neurotoxicity, we propose that a simpler 2D tissue model may prove sufficient. We thus compare the accuracy of predictive models trained on data from a 2D tissue model with those trained on data from a 3D tissue model, and find the 2D model to be substantially more accurate. Furthermore, we find the 2D model to be more robust under stringent gene set selection, whereas the 3D model suffers substantial accuracy degradation. While both approaches have advantages and disadvantages, we propose that our described 2D approach could be a valuable tool for decision makers when prioritizing neurotoxicity screening.

A Simple Text Mining Approach for Ranking Pairwise Associations in Biomedical Applications.

We present a simple text mining method that is easy to implement, requires minimal data collection and preparation, and is easy to use for proposing ranked associations between a list of target terms and a key phrase. We call this method KinderMiner, and apply it to two biomedical applications. The first application is to identify relevant transcription factors for cell reprogramming, and the second is to identify potential drugs for investigation in drug repositioning. We compare the results from our algorithm to existing data and state-of-the-art algorithms, demonstrating compelling results for both application areas. While we apply the algorithm here for biomedical applications, we argue that the method is generalizable to any available corpus of sufficient size.

Leveraging Expert Knowledge to Improve Machine-Learned Decision Support Systems.

While the use of machine learning methods in clinical decision support has great potential for improving patient care, acquiring standardized, complete, and sufficient training data presents a major challenge for methods relying exclusively on machine learning techniques. Domain experts possess knowledge that can address these challenges and guide model development. We present Advice-Based-Learning (ABLe), a framework for incorporating expert clinical knowledge into machine learning models, and show results for an example task: estimating the probability of malignancy following a non-definitive breast core needle biopsy. By applying ABLe to this task, we demonstrate a statistically significant improvement in specificity (24.0% with p=0.004) without missing a single malignancy.

Machine Learning for Treatment Assignment: Improving Individualized Risk Attribution.

Clinical studies model the average treatment effect (ATE), but apply this population-level effect to future individuals. Due to recent developments of machine learning algorithms with useful statistical guarantees, we argue instead for modeling the individualized treatment effect (ITE), which has better applicability to new patients. We compare ATE-estimation using randomized and observational analysis methods against ITE-estimation using machine learning, and describe how the ITE theoretically generalizes to new population distributions, whereas the ATE may not. On a synthetic data set of statin use and myocardial infarction (MI), we show that a learned ITE model improves true ITE estimation and outperforms the ATE. We additionally argue that ITE models should be learned with a consistent, nonparametric algorithm from unweighted examples and show experiments in favor of our argument using our synthetic data model and a real data set of D-penicillamine use for primary biliary cirrhosis.

Support Vector Machines for Differential Prediction.

Machine learning is continually being applied to a growing set of fields, including the social sciences, business, and medicine. Some fields present problems that are not easily addressed using standard machine learning approaches and, in particular, there is growing interest in differential prediction. In this type of task we are interested in producing a classifier that specifically characterizes a subgroup of interest by maximizing the difference in predictive performance for some outcome between subgroups in a population. We discuss adapting maximum margin classifiers for differential prediction. We first introduce multiple approaches that do not affect the key properties of maximum margin classifiers, but which also do not directly attempt to optimize a standard measure of differential prediction. We next propose a model that directly optimizes a standard measure in this field, the uplift measure. We evaluate our models on real data from two medical applications and show excellent results.

Using Machine Learning to Identify Benign Cases with Non-Definitive Biopsy.

When mammography reveals a suspicious finding, a core needle biopsy is usually recommended. In 5% to 15% of these cases, the biopsy diagnosis is non-definitive and a more invasive surgical excisional biopsy is recommended to confirm a diagnosis. The majority of these cases will ultimately be proven benign. The use of excisional biopsy for diagnosis negatively impacts patient quality of life and increases costs to the healthcare system. In this work, we employ a multi-relational machine learning approach to predict when a patient with a non-definitive core needle biopsy diagnosis need not undergo an excisional biopsy procedure because the risk of malignancy is low.

Score As You Lift (SAYL): A Statistical Relational Learning Approach to Uplift Modeling.

We introduce Score As You Lift (SAYL), a novel Statistical Relational Learning (SRL) algorithm, and apply it to an important task in the diagnosis of breast cancer. SAYL combines SRL with the marketing concept of uplift modeling, uses the area under the uplift curve to direct clause construction and final theory evaluation, integrates rule learning and probability assignment, and conditions the addition of each new theory rule to existing ones. Breast cancer, the most common type of cancer among women, is categorized into two subtypes: an earlier in situ stage where cancer cells are still confined, and a subsequent invasive stage. Currently older women with in situ cancer are treated to prevent cancer progression, regardless of the fact that treatment may generate undesirable side-effects, and the woman may die of other causes. Younger women tend to have more aggressive cancers, while older women tend to have more indolent tumors. Therefore older women whose in situ tumors show significant dissimilarity with in situ cancer in younger women are less likely to progress, and can thus be considered for watchful waiting. Motivated by this important problem, this work makes two main contributions. First, we present the first multi-relational uplift modeling system, and introduce, implement and evaluate a novel method to guide search in an SRL framework. Second, we compare our algorithm to previous approaches, and demonstrate that the system can indeed obtain differential rules of interest to an expert on real data, while significantly improving the data uplift.