Identification of Novel Biomarkers for Alzheimer's Disease and Related Dementias Using Unbiased Plasma Proteomics

Alzheimer's disease (AD) and related dementias (ADRD) is a complex disease with multiple pathophysiological drivers that determine clinical symptomology and disease progression. These diseases develop insidiously over time, through many pathways and disease mechanisms and continue to have a huge societal impact for affected individuals and their families. While emerging blood-based biomarkers, such as plasma p-tau181 and p-tau217, accurately detect Alzheimer neuropthology and are associated with faster cognitive decline, the full extension of plasma proteomic changes in ADRD remains unknown. Earlier detection and better classification of the different subtypes may provide opportunities for earlier, more targeted interventions, and perhaps a higher likelihood of successful therapeutic development. In this study, we aim to leverage unbiased mass spectrometry proteomics to identify novel, blood-based biomarkers associated with cognitive decline. 1,786 plasma samples from 1,005 patients were collected over 12 years from partcipants in the Massachusetts Alzheimer's Disease Research Center Longitudinal Cohort Study. Patient metadata includes demographics, final diagnoses, and clinical dementia rating (CDR) scores taken concurrently. The Proteograph™ Product Suite (Seer, Inc.) and liquid-chromatography mass-spectrometry (LC-MS) analysis were used to process the plasma samples in this cohort and generate unbiased proteomics data. Data-independent acquisition (DIA) mass spectrometry results yielded 36,259 peptides and 4,007 protein groups. Linear mixed effects models revealed 138 differentially abundant proteins between AD and healthy controls. Machine learning classification models for AD diagnosis identified potential candidate biomarkers including MBP, BGLAP, and APoD. Cox regression models were created to determine the association of proteins with disease progression and suggest CLNS1A, CRISPLD2, and GOLPH3 as targets of further investigation as potential biomarkers. The Proteograph workflow provided deep, unbiased coverage of the plasma proteome at a speed that enabled a cohort study of almost 1,800 samples, which is the largest, deep, unbiased proteomics study of ADRD conducted to date.

Large-scale sequencing identifies multiple genes and rare variants associated with Crohn's disease susceptibility

Genome-wide association studies (GWASs) have identified hundreds of loci associated with Crohn's disease (CD). However, as with all complex diseases, robust identification of the genes dysregulated by noncoding variants typically driving GWAS discoveries has been challenging. Here, to complement GWASs and better define actionable biological targets, we analyzed sequence data from more than 30,000 patients with CD and 80,000 population controls. We directly implicate ten genes in general onset CD for the first time to our knowledge via association to coding variation, four of which lie within established CD GWAS loci. In nine instances, a single coding variant is significantly associated, and in the tenth, ATG4C, we see additionally a significantly increased burden of very rare coding variants in CD cases. In addition to reiterating the central role of innate and adaptive immune cells as well as autophagy in CD pathogenesis, these newly associated genes highlight the emerging role of mesenchymal cells in the development and maintenance of intestinal inflammation.

Rare and common variant discovery in complex disease: the IBD case study

Complex diseases such as inflammatory bowel disease (IBD), which consists of ulcerative colitis and Crohn's disease, are a significant medical burden-70 000 new cases of IBD are diagnosed in the United States annually. In this review, we examine the history of genetic variant discovery in complex disease with a focus on IBD. We cover methods that have been applied to microsatellite, common variant, targeted resequencing and whole-exome and -genome data, specifically focusing on the progression of technologies towards rare-variant discovery. The inception of these methods combined with better availability of population level variation data has led to rapid discovery of IBD-causative and/or -associated variants at over 200 loci; over time, these methods have grown exponentially in both power and ascertainment to detect rare variation. We highlight rare-variant discoveries critical to the elucidation of the pathogenesis of IBD, including those in NOD2, IL23R, CARD9, RNF186 and ADCY7. We additionally identify the major areas of rare-variant discovery that will evolve in the coming years. A better understanding of the genetic basis of IBD and other complex diseases will lead to improved diagnosis, prognosis, treatment and surveillance.

Automated Classification of Radiographic Knee Osteoarthritis Severity Using Deep Neural Networks

Purpose

To develop an automated model for staging knee osteoarthritis severity from radiographs and to compare its performance to that of musculoskeletal radiologists.

Materials and Methods

Radiographs from the Osteoarthritis Initiative staged by a radiologist committee using the Kellgren-Lawrence (KL) system were used. Before using the images as input to a convolutional neural network model, they were standardized and augmented automatically. The model was trained with 32 116 images, tuned with 4074 images, evaluated with a 4090-image test set, and compared to two individual radiologists using a 50-image test subset. Saliency maps were generated to reveal features used by the model to determine KL grades.

Results

With committee scores used as ground truth, the model had an average F1 score of 0.70 and an accuracy of 0.71 for the full test set. For the 50-image subset, the best individual radiologist had an average F1 score of 0.60 and an accuracy of 0.60; the model had an average F1 score of 0.64 and an accuracy of 0.66. Cohen weighted κ between the committee and model was 0.86, comparable to intraexpert repeatability. Saliency maps identified sites of osteophyte formation as influential to predictions.

Conclusion

An end-to-end interpretable model that takes full radiographs as input and predicts KL scores with state-of-the-art accuracy, performs as well as musculoskeletal radiologists, and does not require manual image preprocessing was developed. Saliency maps suggest the model's predictions were based on clinically relevant information. Supplemental material is available for this article. © RSNA, 2020.