Development and deployment of a histopathology-based deep learning algorithm for patient prescreening in a clinical trial.

Accurate identification of genetic alterations in tumors, such as Fibroblast Growth Factor Receptor, is crucial for treating with targeted therapies; however, molecular testing can delay patient care due to the time and tissue required. Successful development, validation, and deployment of an AI-based, biomarker-detection algorithm could reduce screening cost and accelerate patient recruitment. Here, we develop a deep-learning algorithm using >3000 H&E-stained whole slide images from patients with advanced urothelial cancers, optimized for high sensitivity to avoid ruling out trial-eligible patients. The algorithm is validated on a dataset of 350 patients, achieving an area under the curve of 0.75, specificity of 31.8% at 88.7% sensitivity, and projected 28.7% reduction in molecular testing. We successfully deploy the system in a non-interventional study comprising 89 global study clinical sites and demonstrate its potential to prioritize/deprioritize molecular testing resources and provide substantial cost savings in the drug development and clinical settings.

Accelerating pharmaceutical R&D with a user-friendly AI system for histopathology image analysis.

A system for analysis of histopathology data within a pharmaceutical R&D environment has been developed with the intention of enabling interdisciplinary collaboration. State-of-the-art AI tools have been deployed as easy-to-use self-service modules within an open-source whole slide image viewing platform, so that non-data scientist users (e.g., clinicians) can utilize and evaluate pre-trained algorithms and retrieve quantitative results. The outputs of analysis are automatically cataloged in the database to track data provenance and can be viewed interactively on the slide as annotations or heatmaps. Commonly used models for analysis of whole slide images including segmentation, extraction of hand-engineered features for segmented regions, and slide-level classification using multi-instance learning are included and new models can be added as needed. The source code that supports running inference with these models internally is backed up by a robust CI/CD pipeline to ensure model versioning, robust testing, and seamless deployment of the latest models. Examples of the use of this system in a pharmaceutical development workflow include glomeruli segmentation, enumeration of podocyte count from WT-1 immuno-histochemistry, measurement of beta-1 integrin target engagement from immunofluorescence, digital glomerular phenotyping from periodic acid-Schiff histology, PD-L1 score prediction using multi-instance learning, and the deployment of the open-source Segment Anything model to speed up annotation.

Knowledge synthesis of 100 million biomedical documents augments the deep expression profiling of coronavirus receptors.

The COVID-19 pandemic demands assimilation of all biomedical knowledge to decode mechanisms of pathogenesis. Despite the recent renaissance in neural networks, a platform for the real-time synthesis of the exponentially growing biomedical literature and deep omics insights is unavailable. Here, we present the nferX platform for dynamic inference from over 45 quadrillion possible conceptual associations from unstructured text, and triangulation with insights from single-cell RNA-sequencing, bulk RNA-seq and proteomics from diverse tissue types. A hypothesis-free profiling of ACE2 suggests tongue keratinocytes, olfactory epithelial cells, airway club cells and respiratory ciliated cells as potential reservoirs of the SARS-CoV-2 receptor. We find the gut as the putative hotspot of COVID-19, where a maturation correlated transcriptional signature is shared in small intestine enterocytes among coronavirus receptors (ACE2, DPP4, ANPEP). A holistic data science platform triangulating insights from structured and unstructured data holds potential for accelerating the generation of impactful biological insights and hypotheses.

Fine mapping and subphenotyping implicates ADRA1B gene variants in psoriasis susceptibility in a Chinese population.

AIM: A genomic region on 5q33.3 lies between and encompasses the IL12B and PTTG1 genes, and contains many potential psoriasis causal variants. We aimed to further examine the influence of variants in and around this region. MATERIALS & METHODS: We used least absolute shrinkage and selection operator (LASSO)-based regression analysis to assess independent contributions of 2171 variants to psoriasis susceptibility and tested them for association with different clinical psoriasis subtypes. RESULTS: We found that ADRA1B gene variants contribute to psoriasis in Chinese population. ADRA1B gene variants have a stronger association with moderate-to-severe disease group and an earlier age at onset of psoriasis than IL-12B and PTTG1 variants. CONCLUSION: The association of variants in the ADRA1B gene with psoriasis could explain why variants in the IL-12B, ADRA1B and PTTG1 gene regions are associated with psoriasis.

Pan-cancer analysis reveals technical artifacts in TCGA germline variant calls.

BACKGROUND: Cancer research to date has largely focused on somatically acquired genetic aberrations. In contrast, the degree to which germline, or inherited, variation contributes to tumorigenesis remains unclear, possibly due to a lack of accessible germline variant data. Here we called germline variants on 9618 cases from The Cancer Genome Atlas (TCGA) database representing 31 cancer types. RESULTS: We identified batch effects affecting loss of function (LOF) variant calls that can be traced back to differences in the way the sequence data were generated both within and across cancer types. Overall, LOF indel calls were more sensitive to technical artifacts than LOF Single Nucleotide Variant (SNV) calls. In particular, whole genome amplification of DNA prior to sequencing led to an artificially increased burden of LOF indel calls, which confounded association analyses relating germline variants to tumor type despite stringent indel filtering strategies. The samples affected by these technical artifacts include all acute myeloid leukemia and practically all ovarian cancer samples. CONCLUSIONS: We demonstrate how technical artifacts induced by whole genome amplification of DNA can lead to false positive germline-tumor type associations and suggest TCGA whole genome amplified samples be used with caution. This study draws attention to the need to be sensitive to problems associated with a lack of uniformity in data generation in TCGA data.

Comprehensive analysis of treatment response phenotypes in rheumatoid arthritis for pharmacogenetic studies.

BACKGROUND: An individual patient's response to a particular drug is influenced by multiple factors, which may include genetic predisposition. Pharmacogenetic studies attempt to discover and estimate the contributions of genetic variants to the variability in response to a drug treatment. The task of identifying the genetic contribution is often complicated by response phenotypes that are based on imprecise or subjective clinical observations. Because the success of a pharmacogenetic study depends on the analysis of a heritable phenotype, it is important to identify phenotypes with a significant heritable component to ensure reliable and reproducible results in subsequent genetic association studies. METHODS: We retrospectively analyzed data collected from 436 rheumatoid arthritis patients treated with golimumab during the phase III GO-FURTHER study. We investigated the reliability of several potential response outcomes after golimumab treatment. Using whole-genome sequencing of the clinical trial cohort, we estimated the heritability of each potential outcome measure. We further performed a longitudinal analysis of the clinical data to estimate variability of outcome measures over time and the degree to which each response metric could be confounded by placebo response. RESULTS: We determined that the high degree of within-patient variation over time makes a single follow-up visit insufficient to assess an individual patient's response to golimumab treatment. We found that different potential response outcomes had varying degrees of heritability and that averaging across multiple follow-up visits yielded higher heritability estimates than single follow-up estimates. Importantly, we found that the change in swollen and tender joint counts were the most heritable outcome metrics we tested; however, we showed that they are also more likely to be confounded by a placebo response than objective phenotypes like the change in C-reactive protein levels. CONCLUSIONS: Our rigorous approach to finding robust and heritable response phenotypes could be beneficial to all pharmacogenetic studies and may lead to more reliable and reproducible results. TRIAL REGISTRATION: Clinicaltrials.gov NCT00973479 . Registered 4 September 2009.

Group-based variant calling leveraging next-generation supercomputing for large-scale whole-genome sequencing studies.

MOTIVATION: Next-generation sequencing (NGS) technologies have become much more efficient, allowing whole human genomes to be sequenced faster and cheaper than ever before. However, processing the raw sequence reads associated with NGS technologies requires care and sophistication in order to draw compelling inferences about phenotypic consequences of variation in human genomes. It has been shown that different approaches to variant calling from NGS data can lead to different conclusions. Ensuring appropriate accuracy and quality in variant calling can come at a computational cost. RESULTS: We describe our experience implementing and evaluating a group-based approach to calling variants on large numbers of whole human genomes. We explore the influence of many factors that may impact the accuracy and efficiency of group-based variant calling, including group size, the biogeographical backgrounds of the individuals who have been sequenced, and the computing environment used. We make efficient use of the Gordon supercomputer cluster at the San Diego Supercomputer Center by incorporating job-packing and parallelization considerations into our workflow while calling variants on 437 whole human genomes generated as part of large association study. CONCLUSIONS: We ultimately find that our workflow resulted in high-quality variant calls in a computationally efficient manner. We argue that studies like ours should motivate further investigations combining hardware-oriented advances in computing systems with algorithmic developments to tackle emerging 'big data' problems in biomedical research brought on by the expansion of NGS technologies.