Presenting the Solaris ZFS filesystem, as implemented in Linux FUSE, native kernel modules and the Antergos Linux installer. ZFS remains one of the most technically advanced and feature-complete filesystems since it appeared in October 2005.
Source: ZFS for Linux | Linux Journal
Predictive modeling with electronic health record (EHR) data is anticipated to drive personalized medicine and improve healthcare quality. Constructing predictive statistical models typically requires extraction of curated predictor variables from normalized EHR data, a labor-intensive process that discards the vast majority of information in each patient’s record. We propose a representation of patients’ entire, raw EHR records based on the Fast Healthcare Interoperability Resources (FHIR) format. We demonstrate that deep learning methods using this representation are capable of accurately predicting multiple medical events from multiple centers without site-specific data harmonization. We validated our approach using de-identified EHR data from two U.S. academic medical centers with 216,221 adult patients hospitalized for at least 24 hours. In the sequential format we propose, this volume of EHR data unrolled into a total of 46,864,534,945 data points, including clinical notes. Deep learning models achieved high accuracy for tasks such as predicting in-hospital mortality (AUROC across sites 0.93-0.94), 30-day unplanned readmission (AUROC 0.75-0.76), prolonged length of stay (AUROC 0.85-0.86), and all of a patient’s final discharge diagnoses (frequency-weighted AUROC 0.90). These models outperformed state-of-the-art traditional predictive models in all cases. We also present a case-study of a neural-network attribution system, which illustrates how clinicians can gain some transparency into the predictions. We believe that this approach can be used to create accurate and scalable predictions for a variety of clinical scenarios, complete with explanations that directly highlight evidence in the patient’s chart.
Source: [1801.07860] Scalable and accurate deep learning for electronic health records
A new feature on the iPhone Health app will allow users to automatically download blood test results and other data from their health care providers.
Source: Apple, in Sign of Health Ambitions, Adds Medical Records Feature for iPhone – The New York Times
The technology inside Amazon’s new convenience store, opening Monday in downtown Seattle, enables a shopping experience like no other — including no checkout lines.
Source: Inside Amazon Go, a Store of the Future – The New York Times
The recent generation of gold-standard genome datasets – such as the Genome in a Bottle (GIAB) and Platinum Genomes – has given greater understanding about the quality of sequencing. Competitions held on PrecisionFDA have driven the innovation and refinement of new methods. Sentieon, Edico, and DeepVariant, each a winner of one or more of Consistency, Truth, and Hidden Treasures Challenges, have proven the power of standards to drive innovation.These gold-standard datasets used to construct truth sets and challenges were produced with great care and attention to quality.
We wondered how directly these sets reflect common, real-world sequencing data. “Quality” of sequencing is complex – factors such as library complexity, coverage, contamination, index hopping, PCR errors, and insert size distributionare all important. For our exploration, we use a quantifiable, unambiguous, and straightforward measure – the sequencer-reported base quality.
FASTQ base qualities estimate the probability that a base call (A, T, G, or C) is in error. These errors make reads more difficult to correctly map and produce noise that variant callers must contend with. Monitoring of base qualities is often a component in QC.
We score each pair of forward and reverse reads with the expected value for the number of errors present in the pair. The combination of these values for each read pair in a sequencing run creates a distribution with a mean and standard deviation for errors.
Source: Evaluating the Performance of NGS Pipelines on Noisy WGS Data – Inside DNAnexus