Using Machine Learning to Discover Unexpected Patterns in Clinical Data: A Case Study in COVID-19 Sub-cohort Discovery

As clinicians are faced with a deluge of new information, data science can play a key role in highlighting key features towards developing new clinical hypotheses. Indeed, insights derived from machine learning can serve as a clinical support tool by connecting care providers with results from big data analysis to identify latent patterns that may not be easily detected by even skilled human observers. In this work, we show an example of collaboration between clinicians and data scientists during the COVID-19 pandemic, identifying subgroups of COVID-19 patients with unanticipated outcomes or who are high-risk for severe disease or death. We apply a random forest classifier model to predict adverse patient outcomes early in the disease course, and we connect our classification results to unsupervised clustering of patient features that may underpin patient risk. The paradigm for using data science for hypothesis generation and clinical decision support, as well as our triage classification approach and unsupervised clustering methods to determine patient cohorts, are applicable to driving rapid hypothesis generation and iteration in a variety of clinical challenges, including future public health crises.

GraphOmics: an interactive platform to explore and integrate multi-omics data

Background An increasing number of studies now produce multiple omics measurements that require using sophisticated computational methods for analysis. While each omics data can be examined separately, jointly integrating multiple omics data allows for deeper understanding and insights to be gained from the study. In particular, data integration can be performed horizontally, where biological entities from multiple omics measurements are mapped to common reactions and pathways. However, data integration remains a challenge due to the complexity of the data and the difficulty in interpreting analysis results. Results Here we present GraphOmics, a user-friendly platform to explore and integrate multiple omics datasets and support hypothesis generation. Users can upload transcriptomics, proteomics and metabolomics data to GraphOmics. Relevant entities are connected based on their biochemical relationships, and mapped to reactions and pathways from Reactome. From the Data Browser in GraphOmics, mapped entities and pathways can be ranked, sorted and filtered according to their statistical significance (p values) and fold changes. Context-sensitive panels provide information on the currently selected entities, while interactive heatmaps and clustering functionalities are also available. As a case study, we demonstrated how GraphOmics was used to interactively explore multi-omics data and support hypothesis generation using two complex datasets from existing Zebrafish regeneration and Covid-19 human studies. Conclusions GraphOmics is fully open-sourced and freely accessible from https://graphomics.glasgowcompbio.org/. It can be used to integrate multiple omics data horizontally by mapping entities across omics to reactions and pathways. Our demonstration showed that by using interactive explorations from GraphOmics, interesting insights and biological hypotheses could be rapidly revealed.

PCR, PLS, or OPLS Evaluation of different regression techniques for hypothesis generation

In the current era of ‘big data’, scientists are able to quickly amass enormous amount of data in a limited number of experiments. The investigators then try to hypothesize about the root cause based on the observed trends for the predictors and the response variable. This involves identifying the discriminatory predictors that are most responsible for explaining variation in the response variable. In the current work, we investigated three related multivariate techniques: Principal Component Regression (PCR), Partial Least Squares or Projections to Latent Structures (PLS), and Orthogonal Partial Least Squares (OPLS). To perform a comparative analysis, we used a publicly available dataset for Parkinson’ disease patien ts. We first performed the analysis using a cross-validated number of principal components for the aforementioned techniques. Our results demonstrated that PLS and OPLS were better suited than PCR for identifying the discriminatory predictors. Since the X data did not exhibit a strong correlation, we also performed Multiple Linear Regression (MLR) on the dataset. A comparison of the top five discriminatory predictors identified by the four techniques showed a substantial overlap between the results obtained by PLS, OPLS, and MLR, and the three techniques exhibited a significant divergence from the variables identified by PCR. A further investigation of the data revealed that PCR could be used to identify the discriminatory variables successfully if the number of principal components in the regression model were increased. In summary, we recommend using PLS or OPLS for hypothesis generation and systemizing the selection process for principal components when using PCR.rewordexplain later why MLR can be used on a dataset with no correlation

CAMO: A molecular congruence analysis framework for evaluating model organisms

CAMO provides a rigorous and user-friendly solution for quantification and mechanistic exploration of omics congruence in model organisms and humans. It performs threshold-free differential analysis, quantitative concordance/discordance scoring, pathway-centric investigation, text-mining-based knowledge retrieval, and topological subnetwork detection. Instead of dichotomous claims of "poorly" or "greatly" mimicking, CAMO facilitates discovery and visualization of specific molecular mechanisms that are best or least mimicked, providing foundations for hypothesis generation and subsequent translational investigations.

Reduced Incidence of Long-COVID Symptoms Related to Administration of COVID-19 Vaccines Both Before COVID-19 Diagnosis and Up to 12 Weeks After

Both clinical trials and studies leveraging real-world data have repeatedly confirmed the three COVID-19 vaccines authorized for use by the Food and Drug Administration are safe and effective at preventing infection, hospitalization, and death due to COVID-19 and a recent observational study of self-reported symptoms provides support that vaccination may also reduce the probability of developing long-COVID. As part of a federated research study with the COVID-19 Patient Recovery Alliance, Arcadia.io performed a retrospective analysis of the medical history of 240,648 COVID-19-infected persons to identity factors influencing the development and progression of long-COVID. This analysis revealed that patients who received at least one dose of any of the three COVID vaccines prior to their diagnosis with COVID-19 were 7-10 times less likely to report two or more long-COVID symptoms compared to unvaccinated patients. Furthermore, unvaccinated patients who received their first COVID-19 vaccination within four weeks of SARS-CoV-2 infection were 4-6 times less likely to report multiple long-COVID symptoms, and those who received their first dose 4-8 weeks after diagnosis were 3 times less likely to report multiple long-COVID symptoms compared to those who remained unvaccinated. This relationship supports the hypothesis that COVID-19 vaccination is protective against long-COVID and that effect persists even if vaccination occurs up to 12 weeks after COVID-19 diagnosis. A critical objective of this study was hypothesis generation, and the authors intend to perform further studies to substantiate the findings and encourage other researchers to as well.

Interpretable Visualization of Scientific Hypotheses in Literature-based Discovery

In this paper we present an approach for interpretable visualization of scientific hypotheses that is based on the idea of semantic concept interconnectivity, network-based and topic modeling methods. Our visualization approach has numerous adjustable parameters which provides the domain experts with additional flexibility in their decision making process. We also make use of the Unified Medical Language System metadata by integrating it directly into the resulting topics, and adding the variability into hypotheses resolution. To demonstrate the proposed approach in action, we deployed end-to-end hypothesis generation pipeline AGATHA, which was evaluated by BioCreative VII experts with COVID-19-related queries.

The element of surprise in Peirce’s double consciousness paradigm

This account will demonstrate that the element of surprise is a fundamental device in establishing double consciousness regimes; it further shows how such dialogic paradigms foster abductive inferences by filtering out irrelevant percepts/antecedents. The account sets up Peirce’s Pheme to be the primary device which shocks interpreters’ sensibilities (CP 8.266, 1903) – starting them on a course to question conflicting principles between ego and non-ego (CP 5.53, 1903: CP 8.266). The natural disposition of surprise to instantaneously deliver insight into which antecedents are relevant to vital, anomalous consequences demonstrates its indispensability in generating logical and semiotic advances. For Peirce, vividness, as an element of surprise in Secondness, is largely responsible for the nuts and bolts of how surprise facilitates logic; its means to activate searches for plausible antecedents makes vividness the prime candidate. The success of vividness at turning the mind of interpreters to new ways of explaining the consequence largely hinges upon the external properties of the unexpected consequence – not merely objects’ instantaneity, but their striking characteristics, as well. Even though vividness is external (MS 645, 1909; cf. Atkins, Richard Kenneth. 2018. Charles S. Peirce’s phenomenology: Analysis and consciousness. Oxford: Oxford University Press: 198), it produces internal responses, in the form of a war against feelings (CP 8.330, 1904). In this way, double consciousness paradigms are initiated, which, in turn, call for hypothesis generation.

The Diversity Gap: When Diversity Matters for Knowledge

Can diversity make for better science? Although diversity has ethical and political value, arguments for its epistemic value require a bridge between normative and mechanistic considerations, demonstrating why and how diversity benefits collective intelligence. However, a major hurdle is that the benefits themselves are rather mixed: Quantitative evidence from psychology and behavioral sciences sometimes shows a positive epistemic effect of diversity, but often shows a null effect, or even a negative effect. Here we argue that to make progress with these why and how questions, we need first to rethink when one ought to expect a benefit of cognitive diversity. In doing so, we highlight that the benefits of cognitive diversity are not equally distributed about collective intelligence tasks and are best seen for complex, multistage, creative problem solving, during problem posing and hypothesis generation. Throughout, we additionally outline a series of mechanisms relating diversity and problem complexity, and show how this perspective can inform metascience questions.