Person:

Huang, Jie

Phenotypes are the foundation for clinical and genetic studies of disease risk and outcomes. The growth of biobanks linked to electronic medical record (EMR) data has both facilitated and increased the demand for efficient, accurate, and robust approaches for phenotyping millions of patients. Challenges to phenotyping using EMR data include variation in the accuracy of codes, as well as the high level of manual input required to identify features for the algorithm and to obtain gold standard labels. To address these challenges, we developed PheCAP, a high-throughput semi-supervised phenotyping pipeline. PheCAP begins with data from the EMR, including structured data and information extracted from the narrative notes using natural language processing (NLP). The standardized steps integrate automated procedures reducing the level of manual input, and machine learning approaches for algorithm training. PheCAP itself can be executed in 1-2 days if all data are available; however, the timing is largely dependent on the chart review step which typically requires at least 2 weeks. The final products of PheCAP include a phenotype algorithm, the probability of the phenotype for all patients, and a phenotype classification (yes/no).

In the current precision medicine era, more and more samples get genotyped and sequenced. Both researchers and commercial companies expend significant time and resources to reduce the error rate. However, it has been reported that there is a sample mix-up rate of between 0.1% and 1%, not to mention the possibly higher mix-up rate during the down-stream genetic reporting processes. Even on the low end of this estimate, this translates to a significant number of mislabeled samples, especially over the projected one billion people that will be sequenced within the next decade. Here, we first describe a method to identify a small set of Single nucleotide polymorphisms (SNPs) that can uniquely identify a personal genome, which utilizes allele frequencies of five major continental populations reported in the 1000 genomes project and the ExAC Consortium. To make this panel more informative, we added four SNPs that are commonly used to predict ABO blood type, and another two SNPs that are capable of predicting sex. We then implement a web interface (http://qrcme.tech), nicknamed QRC (for QR code based Concordance check), which is capable of extracting the relevant ID SNPs from a raw genetic data, coding its genotype as a quick response (QR) code, and comparing QR codes to report the concordance of underlying genetic datasets. The resulting 80 fingerprinting SNPs represent a significant decrease in complexity and the number of markers used for genetic data labelling and tracking. Our method and web tool is easily accessible to both researchers and the general public who consider the accuracy of complex genetic data as a prerequisite towards precision medicine.

Background: The relationship between inflammation and blood pressure (BP) has been studied mainly in the general population. In this study, we examined the association between inflammation and BP across a broader range of inflammation observed in rheumatoid arthritis (RA) and non-RA outpatients. Methods: We studied subjects from a tertiary care outpatient population with C-reactive protein (CRP) and BP measured on the same date in 2009–2010; RA outpatients were identified using a validated algorithm. General population data were obtained from the National Health and Nutrition Examination Survey (NHANES) as comparison. To study the cross-sectional association between CRP and BP in the three groups, we constructed a generalized additive model. Longitudinal association between CRP and BP was examined using a repeated-measures linear mixed-effects model in RA outpatients with significant change in inflammation at two consecutive time points. Results: We studied 24,325 subjects from the outpatient population, of whom 1811 had RA, and 5561 were from NHANES. In RA outpatients, we observed a positive relationship between CRP and systolic BP (SBP) at CRP < 6 mg/L and an inverse association at CRP ≥ 6 mg/L. A similar inverse U-shaped relationship was observed in non-RA outpatients. In NHANES, we observed a positive relationship between CRP and SBP as demonstrated by previous studies. Longitudinal analysis in RA showed that every 10 mg/L increase in CRP was associated with a 0.38 mmHg reduction in SBP. Conclusions: Across a broad range of CRP observed in RA and non-RA outpatients, we found an inverse U-shaped relationship between CRP and SBP, highlighting a relationship not previously observed when studying the general population. Electronic supplementary material The online version of this article (10.1186/s13075-018-1597-9) contains supplementary material, which is available to authorized users.