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.

Large-scale in vitro functional testing and novel variant scoring via protein modeling provide insights into alkaline phosphatase activity in hypophosphatasia.

Hypophosphatasia (HPP) is a rare metabolic disorder characterized by low tissue-nonspecific alkaline phosphatase (TNSALP) typically caused by ALPL gene mutations. HPP is heterogeneous, with clinical presentation correlating with residual TNSALP activity and/or dominant-negative effects (DNE). We measured residual activity and DNE for 155 ALPL variants by transient transfection and TNSALP enzymatic activity measurement. Ninety variants showed low residual activity and 24 showed DNE. These results encompass all missense variants with carrier frequencies above 1/25,000 from the Genome Aggregation Database. We used resulting data as a reference to develop a new computational algorithm that scores ALPL missense variants and predicts high/low TNSALP enzymatic activity. Our approach measures the effects of amino acid changes on TNSALP dimer stability with a physics-based implicit solvent energy model. We predict mutation deleteriousness with high specificity, achieving a true-positive rate of 0.63 with false-positive rate of 0, with an area under receiver operating curve (AUC) of 0.9, better than all in silico predictors tested. Combining this algorithm with other in silico approaches can further increase performance, reaching an AUC of 0.94. This study expands our understanding of HPP heterogeneity and genotype/phenotype relationships with the aim of improving clinical ALPL variant interpretation.

mRNA structure regulates protein expression through changes in functional half-life.

Messenger RNAs (mRNAs) encode information in both their primary sequence and their higher order structure. The independent contributions of factors like codon usage and secondary structure to regulating protein expression are difficult to establish as they are often highly correlated in endogenous sequences. Here, we used 2 approaches, global inclusion of modified nucleotides and rational sequence design of exogenously delivered constructs, to understand the role of mRNA secondary structure independent from codon usage. Unexpectedly, highly expressed mRNAs contained a highly structured coding sequence (CDS). Modified nucleotides that stabilize mRNA secondary structure enabled high expression across a wide variety of primary sequences. Using a set of eGFP mRNAs with independently altered codon usage and CDS structure, we find that the structure of the CDS regulates protein expression through changes in functional mRNA half-life (i.e., mRNA being actively translated). This work highlights an underappreciated role of mRNA secondary structure in the regulation of mRNA stability.

Large-scale functional LIPA variant characterization to improve birth prevalence estimates of lysosomal acid lipase deficiency.

Lysosomal acid lipase (LAL) deficiency is an autosomal recessive disorder caused by LIPA gene mutations that disrupt LAL activity. We performed in vitro functional testing of 149 LIPA variants to increase the understanding of the variant effects on LAL deficiency and to improve disease prevalence estimates. Chosen variants had been reported in literature or population databases. Functional testing was done by plasmid transient transfection and LAL activity assessment. We assembled a set of 165 published LAL deficient patient genotypes to evaluate this assay's effectiveness to recapitulate genotype/phenotype relationships. Rapidly progressive LAL deficient patients showed negligible enzymatic activity (<1%), whereas patients with childhood/adult LAL deficiency typically have 1-7% average activity. We benchmarked six in silico variant effect prediction algorithms with these functional data. PolyPhen-2 was shown to have a superior area under the receiver operating curve performance. We used functional data along with Genome Aggregation Database (gnomAD) allele frequencies to estimate LAL deficiency birth prevalence, yielding a range of 3.45-5.97 cases per million births in European-ancestry populations. The low estimate only considers functionally assayed variants in gnomAD. The high estimate computes allele frequencies for variants absent in gnomAD, and uses in silico scores for unassayed variants. Prevalence estimates are lower than previously published, underscoring LAL deficiency's rarity.