Clinical implications of inaccurate potassium determination in hemolyzed pediatric blood specimens.

BACKGROUND: Analysis of whole blood specimens is rapid and saves blood, but hemolysis may go undetected and compromise the accuracy of potassium measurement. We aimed to define the frequency and magnitude of error in whole blood potassium measurement. METHODS: 34 months of whole blood and plasma potassium data were extracted from patients aged less than 2 years at the time of sample acquisition. Hemolysis was detected using the plasma "H index." The magnitude of potassium bias was estimated from the difference between paired whole blood and plasma measurement separated by less than 2 h. RESULTS: 56,000 of the 105,000 data points were from plasma and 20 % of these had significant hemolysis. Rates of hemolysis (nearing 50 %) were greatest in the neonatal nursery. Of 662 proximal whole blood and plasma paired results, 8 % had elevated whole blood potassium with a normal plasma value and 4 % had a normal whole blood potassium with reduced plasma potassium. The bias between whole blood and plasma potassium ranged from -1.0 to 4.0 mmol/L. CONCLUSIONS: The use of whole blood analysis brings with it significant risk for error in potassium measurement. Better tools to detect hemolysis in these types of specimens are indicated.

Working with the Electronic Health Record and Laboratory Information System to Maximize Ordering and Reporting of Molecular Microbiology Results.

Molecular microbiology assays have a higher cost of testing compared to traditional methods and need to be utilized appropriately. Results from these assays may also require interpretation and appropriate follow-up. Electronic tools available in the electronic health record and laboratory information system can be deployed both preanalytically and postanalytically to influence ordering behaviors and positively impact diagnostic stewardship. Next generation technologies, such as machine learning and artificial intelligence, have the potential to expand upon the capabilities currently available and warrant additional study and development but also require regulation around their use in health care.

Automating the Detection of IV Fluid Contamination Using Unsupervised Machine Learning.

BACKGROUND: Intravenous (IV) fluid contamination is a common cause of preanalytical error that can delay or misguide treatment decisions, leading to patient harm. Current approaches for detecting contamination rely on delta checks, which require a prior result, or manual technologist intervention, which is inefficient and vulnerable to human error. Supervised machine learning may provide a means to detect contamination, but its implementation is hindered by its reliance on expert-labeled training data. An automated approach that is accurate, reproducible, and practical is needed. METHODS: A total of 25 747 291 basic metabolic panel (BMP) results from 312 721 patients were obtained from the laboratory information system (LIS). A Uniform Manifold Approximation and Projection (UMAP) model was trained and tested using a combination of real patient data and simulated IV fluid contamination. To provide an objective metric for classification, an "enrichment score" was derived and its performance assessed. Our current workflow was compared to UMAP predictions using expert chart review. RESULTS: UMAP embeddings from real patient results demonstrated outliers suspicious for IV fluid contamination when compared with the simulated contamination's embeddings. At a flag rate of 3 per 1000 results, the positive predictive value (PPV) was adjudicated to be 0.78 from 100 consecutive positive predictions. Of these, 58 were previously undetected by our current clinical workflows, with 49 BMPs displaying a total of 56 critical results. CONCLUSIONS: Accurate and automatable detection of IV fluid contamination in BMP results is achievable without curating expertly labeled training data.

Generalizability of a Machine Learning Model for Improving Utilization of Parathyroid Hormone-Related Peptide Testing across Multiple Clinical Centers.

BACKGROUND: Measuring parathyroid hormone-related peptide (PTHrP) helps diagnose the humoral hypercalcemia of malignancy, but is often ordered for patients with low pretest probability, resulting in poor test utilization. Manual review of results to identify inappropriate PTHrP orders is a cumbersome process. METHODS: Using a dataset of 1330 patients from a single institute, we developed a machine learning (ML) model to predict abnormal PTHrP results. We then evaluated the performance of the model on two external datasets. Different strategies (model transporting, retraining, rebuilding, and fine-tuning) were investigated to improve model generalizability. Maximum mean discrepancy (MMD) was adopted to quantify the shift of data distributions across different datasets. RESULTS: The model achieved an area under the receiver operating characteristic curve (AUROC) of 0.936, and a specificity of 0.842 at 0.900 sensitivity in the development cohort. Directly transporting this model to two external datasets resulted in a deterioration of AUROC to 0.838 and 0.737, with the latter having a larger MMD corresponding to a greater data shift compared to the original dataset. Model rebuilding using site-specific data improved AUROC to 0.891 and 0.837 on the two sites, respectively. When external data is insufficient for retraining, a fine-tuning strategy also improved model utility. CONCLUSIONS: ML offers promise to improve PTHrP test utilization while relieving the burden of manual review. Transporting a ready-made model to external datasets may lead to performance deterioration due to data distribution shift. Model retraining or rebuilding could improve generalizability when there are enough data, and model fine-tuning may be favorable when site-specific data is limited.

A Single Reference Interval for Interpreting Serum Free Light Chains across Patients with Varying Renal Function.

BACKGROUND: Serum free light chain (sFLC) assays are interpreted using a sFLC-ratio-based reference interval (manufacturer's interval) that was defined using a cohort of healthy patients. However, renal impairment elevates the sFLC-ratio, leading to a high false positive rate when using the manufacturer's interval. Prior studies have developed renal-specific reference intervals; however, this approach has not been widely adopted due to practical limitations. Thus, there remains a critical need for a renally robust sFLC interpretation method. METHODS: Retrospective data mining was used to define patient cohorts that reflect the spectrum of renal function seen in clinical practice. Two new reference intervals, one based on the sFLC-ratio and one based on a novel principal component analysis (PCA)-based metric, were developed for the FREELITE assay (Binding Site) on the Roche Cobas c501 instrument (Roche). RESULTS: Compared to the manufacturer's reference interval, both new methods exhibited significantly lower false positive rates and greater robustness to renal function while maintaining equivalent sensitivity for monoclonal gammopathy (MG) diagnosis. While not significantly different, the point estimate for sensitivity was highest for the PCA-based approach. CONCLUSION: Renally robust sFLC interpretation using a single reference interval is possible given a reference cohort that reflects the variation in renal function observed in practice. Further studies are needed to achieve sufficient power and determine if the novel PCA-based metric offers superior sensitivity for MG diagnosis. These new methods offer the practical advantages of not requiring an estimated glomerular filtration rate result or multiple reference intervals, thereby lowering practical barriers to implementation.

Optimizing Equity: Working towards Fair Machine Learning Algorithms in Laboratory Medicine.

BACKGROUND: Methods of machine learning provide opportunities to use real-world data to solve complex problems. Applications of these methods in laboratory medicine promise to increase diagnostic accuracy and streamline laboratory operations leading to improvement in the quality and efficiency of healthcare delivery. However, machine learning models are vulnerable to learning from undesirable patterns in the data that reflect societal biases. As a result, irresponsible application of machine learning may lead to the perpetuation, or even amplification, of existing disparities in healthcare outcomes. CONTENT: In this work, we review what it means for a model to be unfair, discuss the various ways that machine learning models become unfair, and present engineering principles emerging from the field of algorithmic fairness. These materials are presented with a focus on the development of machine learning models in laboratory medicine. SUMMARY: We hope that this work will serve to increase awareness, and stimulate further discussion, of this important issue among laboratorians as the field moves forward with the incorporation of machine learning models into laboratory practice.

Cyclic ADP ribose isomers: Production, chemical structures, and immune signaling.

Cyclic adenosine diphosphate (ADP)-ribose (cADPR) isomers are signaling molecules produced by bacterial and plant Toll/interleukin-1 receptor (TIR) domains via nicotinamide adenine dinucleotide (oxidized form) (NAD+) hydrolysis. We show that v-cADPR (2'cADPR) and v2-cADPR (3'cADPR) isomers are cyclized by O-glycosidic bond formation between the ribose moieties in ADPR. Structures of 2'cADPR-producing TIR domains reveal conformational changes that lead to an active assembly that resembles those of Toll-like receptor adaptor TIR domains. Mutagenesis reveals a conserved tryptophan that is essential for cyclization. We show that 3'cADPR is an activator of ThsA effector proteins from the bacterial antiphage defense system termed Thoeris and a suppressor of plant immunity when produced by the effector HopAM1. Collectively, our results reveal the molecular basis of cADPR isomer production and establish 3'cADPR in bacteria as an antiviral and plant immunity-suppressing signaling molecule.

Defining hierarchical protein interaction networks from spectral analysis of bacterial proteomes.

Cellular behaviors emerge from layers of molecular interactions: proteins interact to form complexes, pathways, and phenotypes. We show that hierarchical networks of protein interactions can be defined from the statistical pattern of proteome variation measured across thousands of diverse bacteria and that these networks reflect the emergence of complex bacterial phenotypes. Our results are validated through gene-set enrichment analysis and comparison to existing experimentally derived databases. We demonstrate the biological utility of our approach by creating a model of motility in Pseudomonas aeruginosa and using it to identify a protein that affects pilus-mediated motility. Our method, SCALES (Spectral Correlation Analysis of Layered Evolutionary Signals), may be useful for interrogating genotype-phenotype relationships in bacteria.

A benzodiazepine activator locks Kv7.1 channels open by electro-mechanical uncoupling.

Loss-of-function mutations in Kv7.1 often lead to long QT syndrome (LQTS), a cardiac repolarization disorder associated with arrhythmia and subsequent sudden cardiac death. The discovery of agonistic IKs modulators may offer a new potential strategy in pharmacological treatment of this disorder. The benzodiazepine derivative (R)-L3 potently activates Kv7.1 channels and shortens action potential duration, thus may represent a starting point for drug development. However, the molecular mechanisms underlying modulation by (R)-L3 are still unknown. By combining alanine scanning mutagenesis, non-canonical amino acid incorporation, voltage-clamp electrophysiology and fluorometry, and in silico protein modelling, we show that (R)-L3 not only stimulates currents by allosteric modulation of the pore domain but also alters the kinetics independently from the pore domain effects. We identify novel (R)-L3-interacting key residues in the lower S4-segment of Kv7.1 and observed an uncoupling of the outer S4 segment with the inner S5, S6 and selectivity filter segments.

Products of gut microbial Toll/interleukin-1 receptor domain NADase activities in gnotobiotic mice and Bangladeshi children with malnutrition.

Perturbed gut microbiome development has been linked to childhood malnutrition. Here, we characterize bacterial Toll/interleukin-1 receptor (TIR) protein domains that metabolize nicotinamide adenine dinucleotide (NAD), a co-enzyme with far-reaching effects on human physiology. A consortium of 26 human gut bacterial strains, representing the diversity of TIRs observed in the microbiome and the NAD hydrolase (NADase) activities of a subset of 152 bacterial TIRs assayed in vitro, was introduced into germ-free mice. Integrating mass spectrometry and microbial RNA sequencing (RNA-seq) with consortium membership manipulation disclosed that a variant of cyclic-ADPR (v-cADPR-x) is a specific product of TIR NADase activity and a prominent, colonization-discriminatory, taxon-specific metabolite. Guided by bioinformatic analyses of biochemically validated TIRs, we find that acute malnutrition is associated with decreased fecal levels of genes encoding TIRs known or predicted to generate v-cADPR-x, as well as decreased levels of the metabolite itself. These results underscore the need to consider microbiome TIR NADases when evaluating NAD metabolism in the human holobiont.

A phytobacterial TIR domain effector manipulates NAD+ to promote virulence.

The Pseudomonas syringae DC3000 type III effector HopAM1 suppresses plant immunity and contains a Toll/interleukin-1 receptor (TIR) domain homologous to immunity-related TIR domains of plant nucleotide-binding leucine-rich repeat receptors that hydrolyze nicotinamide adenine dinucleotide (NAD+ ) and activate immunity. In vitro and in vivo assays were conducted to determine if HopAM1 hydrolyzes NAD+ and if the activity is essential for HopAM1's suppression of plant immunity and contribution to virulence. HPLC and LC-MS were utilized to analyze metabolites produced from NAD+ by HopAM1 in vitro and in both yeast and plants. Agrobacterium-mediated transient expression and in planta inoculation assays were performed to determine HopAM1's intrinsic enzymatic activity and virulence contribution. HopAM1 is catalytically active and hydrolyzes NAD+ to produce nicotinamide and a novel cADPR variant (v2-cADPR). Expression of HopAM1 triggers cell death in yeast and plants dependent on the putative catalytic residue glutamic acid 191 (E191) within the TIR domain. Furthermore, HopAM1's E191 residue is required to suppress both pattern-triggered immunity and effector-triggered immunity and promote P. syringae virulence. HopAM1 manipulates endogenous NAD+ to produce v2-cADPR and promote pathogenesis. This work suggests that HopAM1's TIR domain possesses different catalytic specificity than other TIR domain-containing NAD+ hydrolases and that pathogens exploit this activity to sabotage NAD+ metabolism for immune suppression and virulence.

Using information theory to optimize a diagnostic threshold to match physician-ordering practice.

OBJECTIVE: Clinicians order laboratory tests in an effort to reduce diagnostic or therapeutic uncertainty. Information theory provides the opportunity to quantify the degree to which a test result is expected to reduce diagnostic uncertainty. We sought to apply information theory toward the evaluation and optimization of a diagnostic test threshold and to determine if the results would differ from those of conventional methodologies. We used a heparin/PF4 immunoassay (PF4 ELISA) as a case study. MATERIALS AND METHODS: The laboratory database was queried for PF4 ELISA and serotonin release assay (SRA) results during the study period, with the latter serving as the gold standard for the disease heparin-induced thrombocytopenia (HIT). The optimized diagnostic threshold of the PF4 ELISA test was compared using conventional versus information theoretic approaches under idealized (pretest probability = 50%) and realistic (pretest probability = 2.4%) testing conditions. RESULTS: Under ideal testing conditions, both analyses yielded a similar optimized optical density (OD) threshold of OD > 0.79. Under realistic testing conditions, information theory suggested a higher threshold, OD > 1.5 versus OD > 0.6. Increasing the diagnostic threshold improved the global information value, the value of a positive test and the noise content with only a minute change in the negative test value. DISCUSSION: Our information theoretic approach suggested that the current FDA approved cutoff (OD > 0.4) is overly permissive leading to loss of test value and injection of noise into an already complex diagnostic dilemma. Because our approach is purely statistical and takes as input data that are readily accessible in the clinical laboratory it offers a scalable and data-driven strategy for optimizing test value that may be widely applicable in the domain of laboratory medicine. CONCLUSION: Information theory provides more meaningful measures of test value than the widely used accuracy-based metrics.

Domain-centric database to uncover structure of minimally characterized viral genomes.

Protein domain-based approaches to analyzing sequence data are valuable tools for examining and exploring genomic architecture across genomes of different organisms. Here, we present a complete dataset of domains from the publicly available sequence data of 9,051 reference viral genomes. The data provided contain information such as sequence position and neighboring domains from 30,947 pHMM-identified domains from each reference viral genome. Domains were identified from viral whole-genome sequence using automated profile Hidden Markov Models (pHMM). This study also describes the framework for constructing "domain neighborhoods", as well as the dataset representing it. These data can be used to examine shared and differing domain architectures across viral genomes, to elucidate potential functional properties of genes, and potentially to classify viruses.

Two-stage electro-mechanical coupling of a KV channel in voltage-dependent activation.

In voltage-gated potassium (KV) channels, the voltage-sensing domain (VSD) undergoes sequential activation from the resting state to the intermediate state and activated state to trigger pore opening via electro-mechanical (E-M) coupling. However, the spatial and temporal details underlying E-M coupling remain elusive. Here, utilizing KV7.1's unique two open states, we report a two-stage E-M coupling mechanism in voltage-dependent gating of KV7.1 as triggered by VSD activations to the intermediate and then activated state. When the S4 segment transitions to the intermediate state, the hand-like C-terminus of the VSD-pore linker (S4-S5L) interacts with the pore in the same subunit. When S4 then proceeds to the fully-activated state, the elbow-like hinge between S4 and S4-S5L engages with the pore of the neighboring subunit to activate conductance. This two-stage hand-and-elbow gating mechanism elucidates distinct tissue-specific modulations, pharmacology, and disease pathogenesis of KV7.1, and likely applies to numerous domain-swapped KV channels.

Kinetic Approach Extends the Analytical Measurement Range and Corrects Antigen Excess in Homogeneous Turbidimetric Immunoassays.

BACKGROUND: Homogeneous turbidimetric immunoassays are widely used in the clinical laboratory and offer short assay times, reduced reagent costs, and the potential for full automation. However, these assays have a limited analytical measurement range (AMR) above which antigen excess leads to falsely low estimates of the analyte concentration (i.e., the hook effect). Traditional methods for correction of antigen excess require sample dilution, compromising time and cost-efficiency. Therefore, novel methods that extend the AMR are needed. METHODS: A kinetic model of a generic homogeneous turbidimetric immunoassay was built and then parameterized using a genetic algorithm. Kinetic features that could be used to extend the AMR were identified and subsequently validated with clinical data from consecutive measurements of 2 homogeneous turbidimetric immunoassays: κ serum free light chain and rheumatoid factor. RESULTS: A novel kinetic parameter, the area under the curvature (AUCU), was derived that increases in proportion to the analyte concentration in a range beyond the AMR of conventional end point methods. When applied to clinical data, the AUCU method provided a log-linear calibration curve in the zone of antigen excess extending the AMR by >10-fold for 2 different immunoassays. CONCLUSIONS: The AUCU method detects and corrects antigen excess, extending the AMR in homogeneous turbidimetric immunoassays. The advantage of this method over conventional methods would be a reduction in the number of repeated samples, resulting in significant time and cost savings.