A data-driven architecture using natural language processing to improve phenotyping efficiency and accelerate genetic diagnoses of rare disorders.

Effective genetic diagnosis requires the correlation of genetic variant data with detailed phenotypic information. However, manual encoding of clinical data into machine-readable forms is laborious and subject to observer bias. Natural language processing (NLP) of electronic health records has great potential to enhance reproducibility at scale but suffers from idiosyncrasies in physician notes and other medical records. We developed methods to optimize NLP outputs for automated diagnosis. We filtered NLP-extracted Human Phenotype Ontology (HPO) terms to more closely resemble manually extracted terms and identified filter parameters across a three-dimensional space for optimal gene prioritization. We then developed a tiered pipeline that reduces manual effort by prioritizing smaller subsets of genes to consider for genetic diagnosis. Our filtering pipeline enabled NLP-based extraction of HPO terms to serve as a sufficient replacement for manual extraction in 92% of prospectively evaluated cases. In 75% of cases, the correct causal gene was ranked higher with our applied filters than without any filters. We describe a framework that can maximize the utility of NLP-based phenotype extraction for gene prioritization and diagnosis. The framework is implemented within a cloud-based modular architecture that can be deployed across health and research institutions.

Diagnosis of genetic diseases in seriously ill children by rapid whole-genome sequencing and automated phenotyping and interpretation.

By informing timely targeted treatments, rapid whole-genome sequencing can improve the outcomes of seriously ill children with genetic diseases, particularly infants in neonatal and pediatric intensive care units (ICUs). The need for highly qualified professionals to decipher results, however, precludes widespread implementation. We describe a platform for population-scale, provisional diagnosis of genetic diseases with automated phenotyping and interpretation. Genome sequencing was expedited by bead-based genome library preparation directly from blood samples and sequencing of paired 100-nt reads in 15.5 hours. Clinical natural language processing (CNLP) automatically extracted children's deep phenomes from electronic health records with 80% precision and 93% recall. In 101 children with 105 genetic diseases, a mean of 4.3 CNLP-extracted phenotypic features matched the expected phenotypic features of those diseases, compared with a match of 0.9 phenotypic features used in manual interpretation. We automated provisional diagnosis by combining the ranking of the similarity of a patient's CNLP phenome with respect to the expected phenotypic features of all genetic diseases, together with the ranking of the pathogenicity of all of the patient's genomic variants. Automated, retrospective diagnoses concurred well with expert manual interpretation (97% recall and 99% precision in 95 children with 97 genetic diseases). Prospectively, our platform correctly diagnosed three of seven seriously ill ICU infants (100% precision and recall) with a mean time saving of 22:19 hours. In each case, the diagnosis affected treatment. Genome sequencing with automated phenotyping and interpretation in a median of 20:10 hours may increase adoption in ICUs and, thereby, timely implementation of precise treatments.

The NRP1 migraine risk variant shows evidence of association with menstrual migraine.

BACKGROUND: In 2016, a large meta-analysis brought the number of susceptibility loci for migraine to 38. While sub-type analysis for migraine without aura (MO) and migraine with aura (MA) found some loci showed specificity to MO, the study did not test the loci with respect to other subtypes of migraine. This study aimed to test the hypothesis that single nucleotide polymorphisms (SNPs) robustly associated with migraine are individually or collectively associated with menstrual migraine (MM). METHODS: Genotyping of migraine susceptibility SNPs was conducted using the Agena MassARRAY platform on DNA samples from 235 women diagnosed with menstrual migraine as per International Classification for Headache Disorders II (ICHD-II) criteria and 140 controls. Alternative genotyping methods including restriction fragment length polymorphism, pyrosequencing and Sanger sequencing were used for validation. Statistical analysis was performed using PLINK and SPSS. RESULTS: Genotypes of 34 SNPs were obtained and investigated for their potential association with menstrual migraine. Of these SNPs, rs2506142 located near the neuropilin 1 gene (NRP1), was found to be significantly associated with menstrual migraine (p = 0.003). Genomic risk scores were calculated for all 34 SNPs as well as a subset of 7 SNPs that were nearing individual significance. Overall, this analysis suggested these SNPs to be weakly predictive of MM, but of no prognostic or diagnostic value. CONCLUSIONS: Our results suggest that NRP1 may be important in the etiology of MM. It also suggests some genetic commonality between common migraine subtypes (MA and MO) and MM. The identification of associated SNPs may be the starting point to a better understanding of how genetic factors may contribute to the menstrual migraine sub-type.

Investigation of polymorphisms in genes involved in estrogen metabolism in menstrual migraine.

Migraine is a common, disabling headache disorder, which is influenced by multiple genes and environmental triggers. After puberty, the prevalence of migraine in women is three times higher than in men and >50% of females suffering from migraine report a menstrual association, suggesting hormonal fluctuations can influence the risk of migraine attacks. It has been hypothesized that the drop in estrogen during menses is an important trigger for menstrual migraine. Catechol-O-methyltransferase (COMT) and Cytochrome P450 (CYP) enzymes are involved in estrogen synthesis and metabolism. Functional polymorphisms in these genes can influence estrogen levels and therefore may be associated with risk of menstrual migraine. In this study we investigated four single nucleotide polymorphisms in three genes involved in estrogen metabolism that have been reported to impact enzyme levels or function, in a specific menstrual migraine cohort. 268 menstrual migraine cases and 142 controls were genotyped for rs4680 in COMT (Val158Met), rs4646903 and rs1048943 in CYP1A1 (T3801C and Ile462Val) and rs700519 in CYP19A1 (Cys264Arg). Neither genotype nor allele frequencies for the COMT and CYP SNPs genotyped were found to be significantly different between menstrual migraineurs and controls by chi-square analysis (P>0.05). Therefore we did not find association of functional polymorphisms in the estrogen metabolism genes COMT, CYP1A1 or CYP19A1 with menstrual migraine. Further studies are required to assess whether menstrual migraine is genetically distinct from the common migraine subtypes and identify genes that influence risk.

Genetic association and gene expression studies suggest that genetic variants in the SYNE1 and TNF genes are related to menstrual migraine.

BACKGROUND: Menstrual migraine (MM) encompasses pure menstrual migraine (PMM) and menstrually-related migraine (MRM). This study was aimed at investigating genetic variants that are potentially related to MM, specifically undertaking genotyping and mRNA expression analysis of the ESR1, PGR, SYNE1 and TNF genes in MM cases and non-migraine controls. METHODS: A total of 37 variants distributed across 14 genes were genotyped in 437 DNA samples (282 cases and 155 controls). In addition levels of gene expression were determined in 74 cDNA samples (41 cases and 33 controls). Association and correlation analysis were performed using Plink and RStudio. RESULTS: SNPs rs3093664 and rs9371601 in TNF and SYNE1 genes respectively, were significantly associated with migraine in the MM population (p = 0.008; p = 0.009 respectively). Analysis of qPCR results found no significant difference in levels of gene expression between cases and controls. However, we found a significant correlation between the expression of ESR1 and SYNE1, ESR1 and PGR and TNF and SYNE1 in samples taken during the follicular phase of the menstrual cycle. CONCLUSIONS: Our results show that SNPs rs9371601 and rs3093664 in the SYNE1 and TNF genes respectively, are associated with MM. The present study also provides strong evidence to support the correlation of ESR1, PGR, SYNE1 and TNF gene expression in MM.

Predicting menstrual migraine with a home-use fertility monitor.

A home-use fertility monitor was used to time perimenstrual prophylaxis in 27 women with menstrual or menstrually related migraine. Cycle length variability was mostly caused by follicular phase variability; the postovulatory luteal phase was relatively constant. The monitor accurately identified ovulation in >90% of cycles, enabling prediction of menstruation and precise timing of perimenstrual prophylaxis. Ninety-seven percent of women found the monitor useful in predicting menstrual migraine attacks.