Evaluation and optimization of sequence-based gene regulatory deep learning models.

Neural networks have emerged as immensely powerful tools in predicting functional genomic regions, notably evidenced by recent successes in deciphering gene regulatory logic. However, a systematic evaluation of how model architectures and training strategies impact genomics model performance is lacking. To address this gap, we held a DREAM Challenge where competitors trained models on a dataset of millions of random promoter DNA sequences and corresponding expression levels, experimentally determined in yeast, to best capture the relationship between regulatory DNA and gene expression. For a robust evaluation of the models, we designed a comprehensive suite of benchmarks encompassing various sequence types. While some benchmarks produced similar results across the top-performing models, others differed substantially. All top-performing models used neural networks, but diverged in architectures and novel training strategies, tailored to genomics sequence data. To dissect how architectural and training choices impact performance, we developed the Prix Fixe framework to divide any given model into logically equivalent building blocks. We tested all possible combinations for the top three models and observed performance improvements for each. The DREAM Challenge models not only achieved state-of-the-art results on our comprehensive yeast dataset but also consistently surpassed existing benchmarks on Drosophila and human genomic datasets. Overall, we demonstrate that high-quality gold-standard genomics datasets can drive significant progress in model development.

Text-based predictions of COVID-19 diagnosis from self-reported chemosensory descriptions.

BACKGROUND: There is a prevailing view that humans' capacity to use language to characterize sensations like odors or tastes is poor, providing an unreliable source of information. METHODS: Here, we developed a machine learning method based on Natural Language Processing (NLP) using Large Language Models (LLM) to predict COVID-19 diagnosis solely based on text descriptions of acute changes in chemosensation, i.e., smell, taste and chemesthesis, caused by the disease. The dataset of more than 1500 subjects was obtained from survey responses early in the COVID-19 pandemic, in Spring 2020. RESULTS: When predicting COVID-19 diagnosis, our NLP model performs comparably (AUC ROC ~ 0.65) to models based on self-reported changes in function collected via quantitative rating scales. Further, our NLP model could attribute importance of words when performing the prediction; sentiment and descriptive words such as "smell", "taste", "sense", had strong contributions to the predictions. In addition, adjectives describing specific tastes or smells such as "salty", "sweet", "spicy", and "sour" also contributed considerably to predictions. CONCLUSIONS: Our results show that the description of perceptual symptoms caused by a viral infection can be used to fine-tune an LLM model to correctly predict and interpret the diagnostic status of a subject. In the future, similar models may have utility for patient verbatims from online health portals or electronic health records.

Early in the COVID-19 pandemic, people who were infected with SARS-CoV-2 reported changes in smell and taste. To better study these symptoms of SARS-CoV-2 infections and potentially use them to identify infected patients, a survey was undertaken in various countries asking people about their COVID-19 symptoms. One part of the questionnaire asked people to describe the changes in smell and taste they were experiencing. We developed a computational program that could use these responses to correctly distinguish people that had tested positive for SARS-CoV-2 infection from people without SARS-CoV-2 infection. This approach could allow rapid identification of people infected with SARS-CoV-2 from descriptions of their sensory symptoms and be adapted to identify people infected with other viruses in the future.

Expansive linguistic representations to predict interpretable odor mixture discriminability.

Language is often thought as being poorly adapted to precisely describe or quantify smell and olfactory attributes. In this work, we show that semantic descriptors of odors can be implemented in a model to successfully predict odor mixture discriminability, an olfactory attribute. We achieved this by taking advantage of the structure-to-percept model we previously developed for monomolecular odorants, using chemical descriptors to predict pleasantness, intensity and 19 semantic descriptors such as "fish," "cold," "burnt," "garlic," "grass," and "sweet" for odor mixtures, followed by a metric learning to obtain odor mixture discriminability. Through this expansion of the representation of olfactory mixtures, our Semantic model outperforms state of the art methods by taking advantage of the intermediary semantic representations learned from human perception data to enhance and generalize the odor discriminability/similarity predictions. As 10 of the semantic descriptors were selected to predict discriminability/similarity, our approach meets the need of rapidly obtaining interpretable attributes of odor mixtures as illustrated by the difficulty of finding olfactory metamers. More fundamentally, it also shows that language can be used to establish a metric of discriminability in the everyday olfactory space.

Interpretable Stratification for Chronic Kidney Disease Progression Based on Time to Event Analysis.

In Chronic Kidney Disease (CKD), kidneys are damaged and lose their ability to filter blood, leading to a plethora of health consequences that end up in dialysis. Despite its prevalence, CKD goes often undetected at early stages. In order to better understand disease progression, we stratified patients with CKD by considering the time to dialysis from diagnosis of early CKD (stages 1 or 2). To achieve this, we first reduced the number of clinical features in a predictive time-to-dialysis model and identified the top important features on a cohort of ∼ 40, 000 CKD patients. The extracted features were used to stratify a subpopulation of 3, 522 patients that showed anemia and were prescribed for cardiovascular-related drugs and progressed faster to dialysis. On the other side, clustering patients using conventional clustering methods based on their clinical features did not allow such clear interpretation to identify the main factors for leading fast progression to dialysis. To our knowledge this is the first study extracting interpretable features for stratifying a cohort of early CKD patients using time-to-event analysis which could help prevention and the development of new treatments.

Informing clinical assessment by contextualizing post-hoc explanations of risk prediction models in type-2 diabetes.

Medical experts may use Artificial Intelligence (AI) systems with greater trust if these are supported by 'contextual explanations' that let the practitioner connect system inferences to their context of use. However, their importance in improving model usage and understanding has not been extensively studied. Hence, we consider a comorbidity risk prediction scenario and focus on contexts regarding the patients' clinical state, AI predictions about their risk of complications, and algorithmic explanations supporting the predictions. We explore how relevant information for such dimensions can be extracted from Medical guidelines to answer typical questions from clinical practitioners. We identify this as a question answering (QA) task and employ several state-of-the-art Large Language Models (LLM) to present contexts around risk prediction model inferences and evaluate their acceptability. Finally, we study the benefits of contextual explanations by building an end-to-end AI pipeline including data cohorting, AI risk modeling, post-hoc model explanations, and prototyped a visual dashboard to present the combined insights from different context dimensions and data sources, while predicting and identifying the drivers of risk of Chronic Kidney Disease (CKD) - a common type-2 diabetes (T2DM) comorbidity. All of these steps were performed in deep engagement with medical experts, including a final evaluation of the dashboard results by an expert medical panel. We show that LLMs, in particular BERT and SciBERT, can be readily deployed to extract some relevant explanations to support clinical usage. To understand the value-add of the contextual explanations, the expert panel evaluated these regarding actionable insights in the relevant clinical setting. Overall, our paper is one of the first end-to-end analyses identifying the feasibility and benefits of contextual explanations in a real-world clinical use case. Our findings can help improve clinicians' usage of AI models.

Impact of the COVID-19 pandemic on treatment for mental health needs: a perspective on service use patterns and expenditures from commercial medical claims data.

OBJECTIVE: To examine changes in use patterns, cost of healthcare services before and after the outbreak of the COVID-19 pandemic, and their impacts on expenditures for patients receiving treatment for depression, anxiety, eating disorders, and substance use. METHODS: This cross-sectional study employed statistical tests to analyze claims in MarketScan® Commercial Database in March 2020-February 2021 and quarterly from March 2020 to August 2021, compared to respective pre-pandemic periods. The analysis is based on medical episodes created by the Merative™ Medical Episode Grouper (MEG). MEG is a methodology that groups medical and prescription drug claims to create clinically relevant episodes of care. RESULTS: Comparing year-over-year changes, proportion of patients receiving anxiety treatment among all individuals obtaining healthcare services grew 13.7% in the first year of the pandemic (3/2020-2/2021) versus 10.0% in the year before the pandemic (3/2019-2/2020). This, along with a higher growth in price per episode (5.5% versus 4.3%) resulted in a greater increase in per claimant expenditure ($0.61 versus $0.41 per month). In the same periods, proportion of patients receiving treatment for depression grew 3.7% versus 6.9%, but per claimant expenditure grew by same amount due to an increase in price per episode (4.8%). Proportion of patients receiving treatment for anorexia started to increase 21.1% or more in the fall of 2020. Patient proportion of alcohol use in age group 18-34 decreased 17.9% during the pandemic but price per episode increased 26.3%. Patient proportion of opioid use increased 11.5% in March-May 2020 but decreased or had no significant changes in subsequent periods. CONCLUSIONS: We investigated the changes in use patterns and expenditures of mental health patients before and after the outbreak of the COVID-19 pandemic using claims data in MarketScan®. We found that the changes and their financial impacts vary across mental health conditions, age groups, and periods of the pandemic. Some changes are unexpected from previously reported prevalence increases among the general population and could underlie unmet treatment needs. Therefore, mental health providers should anticipate the use pattern changes in services with similar COVID-19 pandemic disruptions and payers should anticipate cost increases due, in part, to increased price and/or service use.

Human-centered explainability for life sciences, healthcare, and medical informatics.

Rapid advances in artificial intelligence (AI) and availability of biological, medical, and healthcare data have enabled the development of a wide variety of models. Significant success has been achieved in a wide range of fields, such as genomics, protein folding, disease diagnosis, imaging, and clinical tasks. Although widely used, the inherent opacity of deep AI models has brought criticism from the research field and little adoption in clinical practice. Concurrently, there has been a significant amount of research focused on making such methods more interpretable, reviewed here, but inherent critiques of such explainability in AI (XAI), its requirements, and concerns with fairness/robustness have hampered their real-world adoption. We here discuss how user-driven XAI can be made more useful for different healthcare stakeholders through the definition of three key personas-data scientists, clinical researchers, and clinicians-and present an overview of how different XAI approaches can address their needs. For illustration, we also walk through several research and clinical examples that take advantage of XAI open-source tools, including those that help enhance the explanation of the results through visualization. This perspective thus aims to provide a guidance tool for developing explainability solutions for healthcare by empowering both subject matter experts, providing them with a survey of available tools, and explainability developers, by providing examples of how such methods can influence in practice adoption of solutions.

Nicotianagandarela (Solanaceae), a new species of 'tobacco' highly endangered from the Quadrilátero Ferrífero in Brazil.

Nicotianagandarela Augsten & Stehmann (Solanaceae), sp. nov., a small 'tobacco' known only from one locality at Serra do Gandarela, in the state of Minas Gerais, Brazil, is described and illustrated. It is morphologically characterized by its rosulate basal leaves, red corolla with a short tube not inflated at the apex, and the peculiar habitat, a shaded site under a rocky outcrop ledge along a forested stream. Phylogenetic analyses based on a combined dataset of nuclear (ITS) and plastid (ndhF, trnLF, and trnSG) DNA sequences revealed that the species belongs to the Nicotianasect.Alatae and is sister to the clade with the remaining species in the section. A key for the identification of Brazilian species of the section is given. The unusual habitat, the small population size, and the intense pressure of mining activities in the surroundings made the species assessed as Critically Endangered (CR), needing conservation efforts to avoid its extinction.

Impairments in odour detection and hedonic ratings of unpleasant smells in asymptomatic university students as SARS-Cov-2 emerged locally.

Sudden olfactory loss in the absence of concurrent nasal congestion is now a well-recognized symptom of COVID-19. We examined olfaction using standardized objective tests of odour detection, identification and hedonics collected from asymptomatic university students before and as SARS-CoV-2 emerged locally. Olfactory performance of students who were tested when the virus is known to be endemic (n = 22) was compared to students tested in the month prior to viral circulation (n = 25), a normative sample assessed during the previous 4 years (n = 272) and those tested in prior years during the same time period. Analyses showed significantly reduced odour detection for the virus exposed cohort compared to students tested before (t = 2.60; P = .01; d = 0.77; CI 0.17, 1.36) and to the normative sample (D = 0.38; P = .005). Odour identification scores were similar, but the exposed cohort rated odours as less unpleasant (P < .001, CLES = 0.77). Hyposmia increased 4.4-fold for students tested 2 weeks before school closure (N = 22) and increased 13.6-fold for students tested in the final week (N = 11). While the unavailability of COVID-19 testing is a limitation, this naturalistic study demonstrates week-by-week increase in hyposmia in asymptomatic students as a virus was circulating on campus, consistent with increasing airborne viral loads. The specific hedonic deficit in unpleasantness appraisal suggests a deficit in the TAAR olfactory receptor class, which conveys the social salience of odours. Assessment of odour detection and hedonic ratings may aid in early detection of SARS-CoV-2 exposure in asymptomatic and pre-symptomatic persons.

Benchmarked approaches for reconstruction of in vitro cell lineages and in silico models of C. elegans and M. musculus developmental trees.

The recent advent of CRISPR and other molecular tools enabled the reconstruction of cell lineages based on induced DNA mutations and promises to solve the ones of more complex organisms. To date, no lineage reconstruction algorithms have been rigorously examined for their performance and robustness across dataset types and number of cells. To benchmark such methods, we decided to organize a DREAM challenge using in vitro experimental intMEMOIR recordings and in silico data for a C. elegans lineage tree of about 1,000 cells and a Mus musculus tree of 10,000 cells. Some of the 22 approaches submitted had excellent performance, but structural features of the trees prevented optimal reconstructions. Using smaller sub-trees as training sets proved to be a good approach for tuning algorithms to reconstruct larger trees. The simulation and reconstruction methods here generated delineate a potential way forward for solving larger cell lineage trees such as in mouse.

Advances in systems biology modeling: 10 years of crowdsourcing DREAM challenges.

Computational and mathematical models are key to obtain a system-level understanding of biological processes, but their limitations have to be clearly defined to allow their proper application and interpretation. Crowdsourced benchmarks in the form of challenges provide an unbiased assessment of methods, and for the past decade, the Dialogue for Reverse Engineering Assessment and Methods (DREAM) organized more than 15 systems biology challenges. From transcription factor binding to dynamical network models, from signaling networks to gene regulation, from whole-cell models to cell-lineage reconstruction, and from single-cell positioning in a tissue to drug combinations and cell survival, the breadth is broad. To celebrate the 5-year anniversary of Cell Systems, we review the genesis of these systems biology challenges and discuss how interlocking the forward- and reverse-modeling paradigms allows to push the rim of systems biology. This approach will persist for systems levels approaches in biology and medicine.

Gene selection for optimal prediction of cell position in tissues from single-cell transcriptomics data.

Single-cell RNA-sequencing (scRNAseq) technologies are rapidly evolving. Although very informative, in standard scRNAseq experiments, the spatial organization of the cells in the tissue of origin is lost. Conversely, spatial RNA-seq technologies designed to maintain cell localization have limited throughput and gene coverage. Mapping scRNAseq to genes with spatial information increases coverage while providing spatial location. However, methods to perform such mapping have not yet been benchmarked. To fill this gap, we organized the DREAM Single-Cell Transcriptomics challenge focused on the spatial reconstruction of cells from the Drosophila embryo from scRNAseq data, leveraging as silver standard, genes with in situ hybridization data from the Berkeley Drosophila Transcription Network Project reference atlas. The 34 participating teams used diverse algorithms for gene selection and location prediction, while being able to correctly localize clusters of cells. Selection of predictor genes was essential for this task. Predictor genes showed a relatively high expression entropy, high spatial clustering and included prominent developmental genes such as gap and pair-rule genes and tissue markers. Application of the top 10 methods to a zebra fish embryo dataset yielded similar performance and statistical properties of the selected genes than in the Drosophila data. This suggests that methods developed in this challenge are able to extract generalizable properties of genes that are useful to accurately reconstruct the spatial arrangement of cells in tissues.

Reduced odor detection and hedonic changes in asymptomatic university students as SARS-CoV-2 emerged locally.

Aerosol droplets have emerged as the primary mode of SARS-Cov-2 transmission and can be spread by infectious asymptomatic/pre-symptomatic persons rendering indicators of latent viral infection essential. Olfactory impairment is now a recognized symptom of COVID-19 and is rapidly becoming one of the most reliable indicators of the disease. We compared olfaction data from asymptomatic students, who were assessed as SARS-CoV-2 was unknowingly spreading locally, to students tested prior to the arrival of the virus. This study was naturalistic by design as testing occurred in the context of four research studies, all of which used the same inclusion/exclusion criteria and the same protocol to objectively assess odor detection, identification, and hedonics with physiological tests. Data from students (Cohort II; N=22) with probable SARS-CoV-2 exposure were compared to students tested just prior to local virus transmission (Cohort I; N=25), and a normative sample of students assessed over the previous four years (N=272). Students in Cohort II demonstrated significantly reduced odor detection sensitivity compared to students in Cohort I (t=2.60; P=.01; d=0.77; CI, 0.17, 1.36), with a distribution skewed towards reduced detection sensitivity (D=0.38; P=.005). Categorically, the exposed group was significantly more likely to have hyposmia (OR=7.74; CI, 3.1, 19.40), particularly the subgroup assessed in the final week before campus closure (OR=13.61; CI, 3.40, 35.66;). The exposed cohort also rated odors as less unpleasant (P<.001, CLES=0.77). A limitation of our study is that participants were not tested for COVID-19 as testing was unavailable in the area. Objective measures of olfaction may detect olfactory impairment in asymptomatic persons who are otherwise unaware of smell loss. The development of cost-effective, objective olfaction tests that could be self-administered regularly could aid in early detection of SARS-CoV-2 exposure, which is vital to combatting this pandemic.

Mitochondrial origins of fractional control in regulated cell death.

Individual cells in clonal populations often respond differently to environmental changes; for binary phenotypes, such as cell death, this can be measured as a fractional response. These types of responses have been attributed to cell-intrinsic stochastic processes and variable abundances of biochemical constituents, such as proteins, but the influence of organelles is still under investigation. We use the response to TNF-related apoptosis inducing ligand (TRAIL) and a new statistical framework for determining parameter influence on cell-to-cell variability through the inference of variance explained, DEPICTIVE, to demonstrate that variable mitochondria abundance correlates with cell survival and determines the fractional cell death response. By quantitative data analysis and modeling we attribute this effect to variable effective concentrations at the mitochondria surface of the pro-apoptotic proteins Bax/Bak. Further, our study suggests that inhibitors of anti-apoptotic Bcl-2 family proteins, used in cancer treatment, may increase the diversity of cellular responses, enhancing resistance to treatment.

Predicting natural language descriptions of mono-molecular odorants.

There has been recent progress in predicting whether common verbal descriptors such as "fishy", "floral" or "fruity" apply to the smell of odorous molecules. However, accurate predictions have been achieved only for a small number of descriptors. Here, we show that applying natural-language semantic representations on a small set of general olfactory perceptual descriptors allows for the accurate inference of perceptual ratings for mono-molecular odorants over a large and potentially arbitrary set of descriptors. This is noteworthy given that the prevailing view is that humans' capacity to identify or characterize odors by name is poor. We successfully apply our semantics-based approach to predict perceptual ratings with an accuracy higher than 0.5 for up to 70 olfactory perceptual descriptors, a ten-fold increase in the number of descriptors from previous attempts. These results imply that the semantic distance between descriptors defines the equivalent of an odorwheel.

Rethinking cancer: current challenges and opportunities in cancer research.

Cancer therapeutics currently have the lowest clinical trial success rate of all major diseases. Partly as a result of the paucity of successful anti-cancer drugs, cancer will soon be the leading cause of mortality in developed countries. As a disease embedded in the fundamentals of our biology, cancer presents difficult challenges that would benefit from uniting experts from a broad cross-section of related and unrelated fields. Combining extant approaches with novel ones could help in tackling this challenging health problem, enabling the development of therapeutics to stop disease progression and prolong patient lives. This goal provided the inspiration for a recent workshop titled 'Rethinking Cancer', which brought together a group of cancer scientists who work in the academic and pharmaceutical sectors of Europe, America and Asia. In this Editorial, we discuss the main themes emerging from the workshop, with the aim of providing a snapshot of key challenges faced by the cancer research community today. We also outline potential strategies for addressing some of these challenges, from understanding the basic evolution of cancer and improving its early detection to streamlining the thorny process of moving promising drug targets into clinical trials.

Predicting human olfactory perception from chemical features of odor molecules.

It is still not possible to predict whether a given molecule will have a perceived odor or what olfactory percept it will produce. We therefore organized the crowd-sourced DREAM Olfaction Prediction Challenge. Using a large olfactory psychophysical data set, teams developed machine-learning algorithms to predict sensory attributes of molecules based on their chemoinformatic features. The resulting models accurately predicted odor intensity and pleasantness and also successfully predicted 8 among 19 rated semantic descriptors ("garlic," "fish," "sweet," "fruit," "burnt," "spices," "flower," and "sour"). Regularized linear models performed nearly as well as random forest-based ones, with a predictive accuracy that closely approaches a key theoretical limit. These models help to predict the perceptual qualities of virtually any molecule with high accuracy and also reverse-engineer the smell of a molecule.

Crowdsourcing biomedical research: leveraging communities as innovation engines.

The generation of large-scale biomedical data is creating unprecedented opportunities for basic and translational science. Typically, the data producers perform initial analyses, but it is very likely that the most informative methods may reside with other groups. Crowdsourcing the analysis of complex and massive data has emerged as a framework to find robust methodologies. When the crowdsourcing is done in the form of collaborative scientific competitions, known as Challenges, the validation of the methods is inherently addressed. Challenges also encourage open innovation, create collaborative communities to solve diverse and important biomedical problems, and foster the creation and dissemination of well-curated data repositories.