Search NASA⌕ Search

SEARCH · Search NASA

Results for “Logistic Regression”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 163 records · Page 9
All sourcesNASA NTRSDOE OSTI
DOE OSTI2025

Random forest models accurately classify synthetic opioids using high-dimensionality mass spectrometry datasets

Detection of novel threat agents presents several challenges, a principle one being the development of untargeted methods to screen an increasing number of threat chemicals whose exact structures are unknown. With the use of Machine Learning (ML) tools, we can guide the development of analytical methods for broad-spectrum detection of unbounded threat chemical families in complex mixtures. Toward this goal, we used nominal mass and high-resolution mass spectrometry data for hundreds of synthetic opioids and non-opioid compounds. We tested two ML techniques, logistic regression and random forest, to develop models towards a practical, implementable method for opioid detection. We found that of these tested ML methods, random forest models resulted in the highest validation accuracy (95+%) for both nominal mass and high-resolution classification of opioids versus non-opioids, with low false positive and false negative rates. The RF models were then used to successfully predict the classification of 10 compounds—five opioids and five non-opioids not part of the training and validation analysis. This application of ML is a critical step towards the development of field-deployable nominal mass spectrometers with ML-driven analyses for classification of emergent threats.

Arasteh, Kourosh [Lawrence Livermore National Labo↗
DOE OSTI2023

COVID-19 Outcomes by Cancer Status, Site, Treatment, and Vaccination

Background: Studies have shown an increased risk of severe SARS-CoV-2–related (COVID-19) disease outcome and mortality for patients with cancer, but it is not well understood whether associations vary by cancer site, cancer treatment, and vaccination status. Methods: Using electronic health record data from an academic medical center, we identified a retrospective cohort of 260,757 individuals tested for or diagnosed with COVID-19 from March 10, 2020, to August 1, 2022. Of these, 52,019 tested positive for COVID-19 of whom 13,752 had a cancer diagnosis. We conducted Firth-corrected logistic regression to assess the association between cancer status, site, treatment, vaccination, and four COVID-19 outcomes: hospitalization, intensive care unit admission, mortality, and a composite “severe COVID” outcome. Results: Cancer diagnosis was significantly associated with higher rates of severe COVID, hospitalization, and mortality. These associations were driven by patients whose most recent initial cancer diagnosis was within the past 3 years. Chemotherapy receipt, colorectal cancer, hematologic malignancies, kidney cancer, and lung cancer were significantly associated with higher rates of worse COVID-19 outcomes. Vaccinations were significantly associated with lower rates of worse COVID-19 outcomes regardless of cancer status. Conclusions: Patients with colorectal cancer, hematologic malignancies, kidney cancer, or lung cancer or who receive chemotherapy for treatment should be cautious because of their increased risk of worse COVID-19 outcomes, even after vaccination. Impact: Additional COVID-19 precautions are warranted for people with certain cancer types and treatments. Further, significant benefit from vaccination is noted for both cancer and cancer-free patients.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2022

Understanding Predictability of Daily Southeast U.S. Precipitation Using Explainable Machine Learning

Abstract We investigate the predictability of the sign of daily southeastern U.S. (SEUS) precipitation anomalies associated with simultaneous predictors of large-scale climate variability using machine learning models. Models using index-based climate predictors and gridded fields of large-scale circulation as predictors are utilized. Logistic regression (LR) and fully connected neural networks using indices of climate phenomena as predictors produce neither accurate nor reliable predictions, indicating that the indices themselves are not good predictors. Using gridded fields as predictors, an LR and convolutional neural network (CNN) are more accurate than the index-based models. However, only the CNN can produce reliable predictions that can be used to identify forecasts of opportunity. Using explainable machine learning we identify which variables and grid points of the input fields are most relevant for confident and correct predictions in the CNN. Our results show that the local circulation is most important as represented by maximum relevance of 850-hPa geopotential heights and zonal winds to making skillful, high-probability predictions. Corresponding composite anomalies identify connections with El Niño–Southern Oscillation during winter and the Atlantic multidecadal oscillation and North Atlantic subtropical high during summer.

Pegion, Kathy↗
DOE OSTI2025

The Effect of Updraft Entrainment on Convective Cell Deepening in Realistic Large-Eddy Simulations

Entrainment of surrounding cooler and drier air into convective updrafts is one of the key processes that influence deep convection initiation and growth. Numerous studies have investigated the effect of entrainment on isolated convective cloud growth in idealized simulations, but the importance of this effect in realistic conditions with many interacting convective clouds remains uncertain. We examine the impact of entrainment on the depth reached by convective clouds in realistic large-eddy simulations (LES) over central Argentina during the Cloud, Aerosol, and Complex Terrain Interactions (CACTI) field campaign. Cloudy updrafts and their associated properties are assigned to convective cells tracked with radar reflectivity signatures. Several thousand convective cells are tracked over two high convective available potential energy (CAPE) and two low CAPE cases that support cells of varying depths and intensities. Entrainment is calculated explicitly as the fluxes of air into the outer surface of each cloudy updraft. Single-predictor logistic regression models are used to determine the relative importance of updraft, near-updraft, and preconvective initiation atmospheric conditions in predicting whether convective cells become deep. We then build a multiple-predictor regression model pairing important updraft and meteorological metrics with fractional entrainment rate. The probability of cells transitioning to deep convection is most sensitive to ambient 600-hPa relative humidity (42% of total metric contribution to cloud depth predictability), followed by low-level CAPE (28%), cloud-base updraft width (19%), and fractional entrainment (11%). Thus, the initial width of the updraft along with potential buoyancy and its dilution through the midtroposphere collectively determine whether deep convection will result from shallower clouds.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2023

Predicting Postoperative Injury and Military Discharge Status After Knee Surgery in the US Army

Background: Researchers have assessed postoperative injury or disability predictors in the military setting but typically focused on 1 type of surgical procedure at a time, used relatively small sample sizes, or investigated mixed cohorts with civilian populations. Purpose: To identify the relationship between baseline variables and injury incidence or military discharge status in US Army soldiers after knee surgery. Study Design: Case-control study; Level of evidence, 3. Methods: Data were obtained from a repository containing personnel, performance, and medical records for all active-duty US Army soldiers. Multivariate logistic regressions were used to estimate the effects of numerous variables on postoperative injury or on medical discharge. Variable selection and model validation were conducted using the k-fold method. Results: A total of 7567 soldiers underwent knee surgery between 2017 and 2019. Meniscal procedures were the most common type of surgery (39%), and approximately 71% of the cohort had a postoperative injury. Significant predictors for sustaining a postoperative injury included having a previous nonknee injury (odds ratio [OR], 1.5), female sex (OR, 1.3), and Black race (OR, 1.2). Within 4 years after surgery, 17% of soldiers were discharged from the military because of knee-related disability. Significant predictors for discharge from duty included enlisted rank (OR, 2.3), recent fitness test failure (OR, 1.9), number of previous knee surgeries (OR, 1.7), and having a previous nonknee injury (OR, 1.6). Conclusion: After knee surgery, nearly three-fourths of the soldiers in this cohort sustained a postoperative injury and almost one-fifth of soldiers were medically discharged from the military within 4 years. This study identified variables that indicate statistically increased risk for these postoperative outcomes and highlighted potentially modifiable factors.

Orthopedics↗
DOE OSTI2023

Contraceptive Sabotage and Contraceptive Use at the Time of Pregnancy: An Analysis of People with a Recent Live Birth in the United States

Contraceptive sabotage and other forms of intimate partner violence (IPV) can interfere with contraceptive use. We used 2012 to 2015 Pregnancy Risk Assessment Monitoring System data from 8,981 people residing in five states who reported that when they became pregnant, they were not trying to get pregnant. We assessed the relationships between ever experiencing contraceptive sabotage and physical IPV 12 months before pregnancy (both by the current partner) and contraceptive use at the time of pregnancy using multivariable logistic regression. We also assessed the joint associations between physical IPV 12 months before pregnancy and ever experienced contraceptive sabotage with contraceptive use at the time of pregnancy. Few people ever experienced contraceptive sabotage (1.8%; 95% confidence interval [CI]: 1.4, 2.3) or physical IPV 12 months before pregnancy (2.8%; 95% CI: 2.3, 3.3). In models adjusted for age, race/ethnicity, marital status, education, and state of residence, ever experiencing contraceptive sabotage was associated with contraceptive use at the time of pregnancy (adjusted odds ratio [aOR]: 1.73; 95% CI: 1.06, 2.82), but not with physical IPV 12 months before pregnancy (aOR: 0.69; 95% CI: 0.46, 1.02). When examining the joint association, compared to not ever experiencing contraceptive sabotage or physical IPV 12 months before pregnancy, ever experiencing contraceptive sabotage was significantly related to contraceptive use at the time of pregnancy (aOR: 1.72; 95% CI: 1.00, 2.95). However, it was not associated with experiencing physical IPV 12 months before pregnancy (aOR: 0.68; 95% CI: 0.45, 1.04) or with experiencing both contraceptive sabotage and physical IPV 12 months before pregnancy (aOR: 1.21; 95% CI: 0.42, 3.50), compared to not ever experiencing contraceptive sabotage or physical IPV 12 months before pregnancy. Our study highlights that current partner contraceptive sabotage may motivate those not trying to get pregnant to use contraception; however, all people in our sample still experienced a pregnancy.

Huber-Krum, Sarah↗
DOE OSTI2022

Investigation of Plasmodium falciparum pfhrp2 and pfhrp3 gene deletions and performance of a rapid diagnostic test for identifying asymptomatic malaria infection in northern Ethiopia, 2015

Background: Rapid diagnostic tests (RDTs) are widely used for malaria diagnosis of both symptomatic and asymptomatic infections. Although RDTs are a reliable and practical diagnostic tool, the sensitivity of histidine-rich protein 2 (HRP2)-based RDTs can be reduced if pfhrp2 or pfhrp3 (pfhrp2/3) gene deletions exist in the Plasmodium falciparum parasite population. This study evaluated dried blood spot (DBS) samples collected from a national household survey to investigate the presence of pfhrp2/3 deletions and the performance of the RDT used in the cross-sectional survey in a low transmission setting. Methods: The 2015 Ethiopia Malaria Indicator Survey tested household members by RDT and collected DBS samples. DBS (n = 2648) from three regions in northern Ethiopia were tested by multiplex bead-based antigen detection assay after completion of the survey. The multiplex assay detected pan-Plasmodium lactate dehydrogenase (LDH), pAldolase, and HRP2 antigens in samples. Samples suspected for pfhrp2/3 gene deletions (pLDH and/or pAldolase positive but low or absent HRP2) were further investigated by molecular assays for gene deletions. Antigen results were also compared to each individual’s RDT results. Dose–response logistic regression models were fit to estimate RDT level of detection (LOD) antigen concentrations at which 50, 75, 90, and 95% of the RDTs returned a positive result during this survey. Results: Out of 2,648 samples assayed, 29 were positive for pLDH or pAldolase antigens but low or absent for HRP2 signal, and 15 of these samples (51.7%) were successfully genotyped for pfhrp2/3. Of these 15 P. falciparum infections, eight showed single deletions in pfhrp3, one showed a single pfhrp2 deletion, and six were pfhrp2/3 double-deletions. Six pfhrp2 deletions were observed in Tigray and one in Amhara. Twenty-five were positive for HRP2 by the survey RDT while the more sensitive bead assay detected 30 HRP2-positive samples. A lower concentration of HRP2 antigen generated a positive test result by RDT compared to pLDH (95% LOD: 16.9 ng/mL vs. 319.2 ng/mL, respectively). Conclusions: There is evidence of dual pfhrp2/3 gene deletions in the Tigray and Amhara regions of Ethiopia in 2015. As the prevalence of malaria was very low (< 2%), it is difficult to make strong conclusions on RDT performance, but these results challenge the utility of biomarkers in household surveys in very low transmission settings.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2024

Organochlorine pesticides and risk of papillary thyroid cancer in U.S. military personnel: a nested case-control study

Background: The effects of organochlorine pesticide (OCP) exposure on the development of human papillary thyroid cancer (PTC) are not well understood. A nested case-control study was conducted with data from the U.S. Department of Defense Serum Repository (DoDSR) cohort between 2000 and 2013 to assess associations of individual OCPs serum concentrations with PTC risk. Methods: This study included 742 histologically confirmed PTC cases (341 females, 401 males) and 742 individually-matched controls with pre-diagnostic serum samples selected from the DoDSR. Associations between categories of lipid-corrected serum concentrations of seven OCPs and PTC risk were evaluated for classical PTC and follicular PTC using conditional logistic regression, adjusted for body mass index category and military branch to compute odds ratios (OR) and 95% confidence intervals (CIs). Effect modification by sex, birth cohort, and race was examined. Results: There was no evidence of associations between most of the OCPs and PTC, overall or stratified by histological subtype. Overall, there was no evidence of an association between hexachlorobenzene (HCB) and PTC, but stratified by histological subtype HCB was associated with significantly increased risk of classical PTC (third tertile above the limit of detection (LOD) vs. <LOD, OR = 1.61, 95% CI, 1.09, 2.38; p for trend = 0.05) and significantly decreased risk of follicular variant PTC (third tertile above the limit of detection (LOD) vs. <LOD, OR = 0.38, 95% CI, 0.16, 0.91; p for trend = 0.04). Further stratified by sex, risk of classical PTC was higher for females (third tertile above LOD vs. <LOD, OR = 2.23, 95% CI: 1.23, 4.06; p-trend = 0.02) than for males (OR = 1.22, 95%CI: 0.72–2.08; p-trend = 0.56), though the test for interaction by sex was not statistically significant (p-interaction = 0.30). Similarly, β-hexachlorocyclohexane (β-HCCH) was associated with a higher risk for classical PTC for women with concentrations ≥LOD versus <LOD (OR = 1.76, 95% CI: 1.07, 2.89), while the effects were null for men. There were no consistent trends when stratified by race or birth year. Conclusions: The U.S. Environmental Protection Agency has classified HCB and other OCPs we studied here as probable human carcinogens. Our findings of increased risks for classical PTC associated with increased concentrations of HCB and β-HCCH, which were stronger among females, should be replicated in future studies of other populations.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2025

The impact of kidney function on Alzheimer’s disease blood biomarkers: implications for predicting amyloid-β positivity

Impaired kidney function has a potential confounding effect on blood biomarker levels, including biomarkers for Alzheimer’s disease (AD). Given the imminent use of certain blood biomarkers in the routine diagnostic work-up of patients with suspected AD, knowledge on the potential impact of comorbidities on the utility of blood biomarkers is important. We aimed to evaluate the association between kidney function, assessed through estimated glomerular filtration rate (eGFR) calculated from plasma creatinine and AD blood biomarkers, as well as their influence over predicting Aβ-positivity. We included 242 participants from the Translational Biomarkers in Aging and Dementia (TRIAD) cohort, comprising cognitively unimpaired individuals (CU; n = 124), mild cognitive impairment (MCI; n = 58), AD dementia (n = 34), and non-AD dementia (n = 26) patients all characterized by [ 18 F] AZD-4694. Plasma samples were analyzed for Aβ42, Aβ40, glial fibrillary acidic protein (GFAP), neurofilament light chain (NfL), tau phosphorylated at threonine 181 (p-tau181), 217 (p-tau217), 231 (p-tau231) and N-terminal containing tau fragments (NTA-tau) using Simoa technology. Kidney function was assessed by eGFR in mL/min/1.73 m 2 , based on plasma creatinine levels, age, and sex. Participants were also stratified according to their eGFR-indexed stages of chronic kidney disease (CKD). We evaluated the association between eGFR and blood biomarker levels with linear models and assessed whether eGFR provided added predictive value to determine Aβ-positivity with logistic regression models. Biomarker concentrations were highest in individuals with CKD stage 3, followed by stages 2 and 1, but differences were only significant for NfL, Aβ42, and Aβ40 (not Aβ42/Aβ40). All investigated biomarkers showed significant associations with eGFR except plasma NTA-tau, with stronger relationships observed for Aβ40 and NfL. However, after adjusting for either age, sex or Aβ-PET SUVr, the association with eGFR was no longer significant for all biomarkers except Aβ40, Aβ42, NfL, and GFAP. When evaluating whether accounting for kidney function could lead to improved prediction of Aβ-positivity, we observed no improvements in model fit (Akaike Information Criterion, AIC) or in discriminative performance (AUC) by adding eGFR to a base model including each plasma biomarker, age, and sex. While covariates like age and sex improved model fit, eGFR contributed minimally, and there were no significant differences in clinical discrimination based on AUC values. We found that kidney function seems to be associated with AD blood biomarker concentrations. However, these associations did not remain significant after adjusting for age and sex, except for Aβ40, Aβ42, NfL, and GFAP. While covariates such as age and sex improved prediction of Aβ-positivity, including eGFR in the models did not lead to improved prediction for any biomarker. Our findings indicate that renal function, within the normal to mild impairment range, does not seem to have a clinically relevant impact when using highly accurate blood biomarkers, such as p-tau217, in a biomarker-supported diagnosis.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2023

The Cross-sectional and Longitudinal Association Between Thyroid Function and Depression: A Population-Based Study

Abstract Context An association of thyroid function with mood disorders has been widely suggested, but very few studies have examined this association longitudinally. Objective We assessed the cross-sectional and longitudinal association between thyroid function and depression in a population-based cohort. Methods A total of 9471 individuals were included in cross-sectional analyses, of whom 8366 had longitudinal data. At baseline, we assessed thyroid function using serum samples (thyrotropin [TSH], free thyroxine (FT4), and thyroid peroxidase antibodies) and depressive symptoms using the Centre for Epidemiologic Studies Depression (CES-D) scale. Incident depressive events (n = 1366) were continuously followed up with the CES-D and clinical interviews. We analyzed the cross-sectional association of thyroid function and thyroid disease with depressive symptoms using linear and logistic regression, and the longitudinal association with Cox proportional hazard models for depressive events. Results Lower TSH levels and lower and higher FT4 levels were cross-sectionally associated with more depressive symptoms with a B value of −0.07 per 1 unit increase of natural log-transformed TSH (95% CI −0.11; −0.04). Furthermore, hypothyroidism was cross-sectionally associated with less depressive symptoms and hyperthyroidism with more depressive symptoms. Longitudinally, there was a U-shaped association between FT4 and incident depressive events but only in euthyroid participants. Conclusion We show a cross-sectional association between thyroid (dys)function with depressive symptoms, and a U-shaped association between FT4 and incident depressive events in euthyroid individuals. Our findings suggest an association of thyroid function with the risk of developing depression, albeit small. Reverse causation and additional underlying factors may also contribute to the association.

Roa Dueñas, Oscar Hernando (ORCID:0000000207522261↗
DOE OSTI2019

Does Breast Density Increase the Risk of Re-excision for Women with Breast Cancer Having Breast-Conservation Therapy?

Background: Women with dense breasts may have less-accurate preoperative evaluation of extent of disease, potentially affecting the achievement of negative margins. The goal of this study is to examine the association between breast density and re-excision rates in women having breast-conserving surgery for invasive breast cancer. Patients and Methods: Women with stage I/II invasive breast cancer treated with breast-conserving surgery between 1/1/2014 and 10/31/2014 were included. Breast density was assessed by two radiologists. The association between breast density and re-excision was examined using logistic regression. Results: Seven hundred and one women were included. Overall, 106 (15.1%) women had at least one re-excision. Younger age at diagnosis was associated with increased breast density (p < 0.001). On univariable analysis, increased breast density was associated with significantly increased odds of re-excision (odds ratio [OR] 1.38, 95% confidence interval [CI] 1.04–1.83), as was multifocal disease, human epidermal growth factor receptor 2 (HER2) positive status, and extensive intraductal component (EIC) (all p < 0.05). On multivariable analysis, breast density remained significantly associated with increased odds of re-excision (OR 1.37, 95% CI 1.00–1.86), as did multifocality and EIC. HER2 positive status was not significantly associated with re-excision on multivariable analysis. Conclusions: Women with dense breasts are more likely to need additional surgery (re-excision after breast-conserving surgery), but increased breast density did not adversely affect disease-free survival in our study. Our findings support the need for further study in developing techniques that can help decrease re-excisions for women with dense breasts who undergo breast-conserving surgery.

62 RADIOLOGY AND NUCLEAR MEDICINE↗
DOE OSTI2019

Trends in Regional Nodal Management of Breast Cancer Patients with Low Nodal Burden

Background: American College of Surgeons Oncology Group Z0011 confirms the safety of omitting axillary lymph node dissection (ALND) and regional nodal irradiation (RNI) in breast cancer patients with one to two positive sentinel lymph nodes (SLNs), without compromising disease-free survival (DFS) and overall survival (OS). In contrast, the NCIC MA20 trial showed improved DFS in node-positive patients undergoing ALND and RNI. We sought to examine how these data have influenced the management of patients with limited nodal burden. Methods: Using the National Cancer Database, patients diagnosed between 2010 and 2015 and who met the criteria for Z0011 were identified. Logistic regression was used to analyze factors associated with practice patterns. The Cox proportional hazards model was used to assess the association of ALND and RNI with OS. Results: Omission of ALND in Z0011-eligible patients reached 89.2% in 2015. This Z0011-compliant group was more likely to undergo RNI compared with the non-compliant group (36.4% vs. 31.3%; p < 0.05), with RNI increasing to 43.8% by 2015. Factors associated with the use of RNI included later year of diagnosis [odds ratio (OR) 1.8, 95% confidence interval (CI) 1.6–2.1], hormone receptor-negative tumor (OR 1.2, 95% CI 1.1–1.4), grade 3 tumor (OR 1.2, 95% CI 1.1–1.3), treatment at a non-academic site (OR 1.2, 95% CI 1.1–1.3) and two versus one positive SLN (OR 2.0, 95% CI 1.8–2.2). With 43 months median follow-up, RNI was not associated with improved OS. Conclusion: Since the publication of Z0011, the omission of ALND has become widespread; however nearly half of these women now receive RNI. The optimal radiation therapy approach for this low nodal burden population warrants further study.

62 RADIOLOGY AND NUCLEAR MEDICINE↗
DOE OSTI2020

Early prediction of antigenic transitions for influenza A/H3N2

Influenza A/H3N2 is a rapidly evolving virus which experiences major antigenic transitions every two to eight years. Anticipating the timing and outcome of transitions is critical to developing effective seasonal influenza vaccines. Using a published phylodynamic model of influenza transmission, we identified indicators of future evolutionary success for an emerging antigenic cluster and quantified fundamental trade-offs in our ability to make such predictions. The eventual fate of a new cluster depends on its initial epidemiological growth rate––which is a function of mutational load and population susceptibility to the cluster––along with the variance in growth rate across co-circulating viruses. Logistic regression can predict whether a cluster at 5% relative frequency will eventually succeed with ~80% sensitivity, providing up to eight months advance warning. As a cluster expands, the predictions improve while the lead-time for vaccine development and other interventions decreases. However, attempts to make comparable predictions from 12 years of empirical influenza surveillance data, which are far sparser and more coarse-grained, achieve only 56% sensitivity. By expanding influenza surveillance to obtain more granular estimates of the frequencies of and population-wide susceptibility to emerging viruses, we can better anticipate major antigenic transitions. This provides added incentives for accelerating the vaccine production cycle to reduce the lead time required for strain selection.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2020

Development and validation of a 30-day mortality index based on pre-existing medical administrative data from 13,323 COVID-19 patients: The Veterans Health Administration COVID-19 (VACO) Index

Background Available COVID-19 mortality indices are limited to acute inpatient data. Using nationwide medical administrative data available prior to SARS-CoV-2 infection from the US Veterans Health Administration (VA), we developed the VA COVID-19 (VACO) 30-day mortality index and validated the index in two independent, prospective samples. Methods and findings We reviewed SARS-CoV-2 testing results within the VA between February 8 and August 18, 2020. The sample was split into a development cohort (test positive between March 2 and April 15, 2020), an early validation cohort (test positive between April 16 and May 18, 2020), and a late validation cohort (test positive between May 19 and July 19, 2020). Our logistic regression model in the development cohort considered demographics (age, sex, race/ethnicity), and pre-existing medical conditions and the Charlson Comorbidity Index (CCI) derived from ICD-10 diagnosis codes. Weights were fixed to create the VACO Index that was then validated by comparing area under receiver operating characteristic curves (AUC) in the early and late validation cohorts and among important validation cohort subgroups defined by sex, race/ethnicity, and geographic region. We also evaluated calibration curves and the range of predictions generated within age categories. 13,323 individuals tested positive for SARS-CoV-2 (median age: 63 years; 91% male; 42% non-Hispanic Black). We observed 480/3,681 (13%) deaths in development, 253/2,151 (12%) deaths in the early validation cohort, and 403/7,491 (5%) deaths in the late validation cohort. Age, multimorbidity described with CCI, and a history of myocardial infarction or peripheral vascular disease were independently associated with mortality–no other individual comorbid diagnosis provided additional information. The VACO Index discriminated mortality in development (AUC = 0.79, 95% CI: 0.77–0.81), and in early (AUC = 0.81 95% CI: 0.78–0.83) and late (AUC = 0.84, 95% CI: 0.78–0.86) validation. The VACO Index allows personalized estimates of 30-day mortality after COVID-19 infection. For example, among those aged 60–64 years, overall mortality was estimated at 9% (95% CI: 6–11%). The Index further discriminated risk in this age stratum from 4% (95% CI: 3–7%) to 21% (95% CI: 12–31%), depending on sex and comorbid disease. Conclusion Prior to infection, demographics and comorbid conditions can discriminate COVID-19 mortality risk overall and within age strata. The VACO Index reproducibly identified individuals at substantial risk of COVID-19 mortality who might consider continuing social distancing, despite relaxed state and local guidelines.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2021

Development and validation of a diabetes risk score among two populations

The purpose of this study was to assess the validity of a practical diabetes risk score amongst two heterogenous populations, a working population and a non-working population. Study population 1 (n = 2,089) participated in a large-scale screening program offered to retired workers to discover previously undetected/incipient chronic illness. Study population 2 (n = 3,293) was part of a Colorado worksite wellness program health risk assessment. We assessed the relationship between a continuous diabetes risk score at baseline and development of diabetes in the future using logistic regression. Receiver operating curves and sensitivity/specificity of the models were calculated. Across both study populations, we observed that participants with diabetes at follow-up had higher diabetes risk scores at baseline than participants who did not have diabetes at follow-up. On average, the odds ratio of developing diabetes in the future was 1.38 (95% CI: 1.26–1.50, p < 0.0001) for study population 1 and 1.68 (95% CI: 1.45–1.95, p-value < 0.0001) for study population 2. These findings indicate that the diabetes risk score may be generalizable to diverse individuals, and thus potentially a population level diabetes screening tool. Minimally-invasive diabetes risk scores can aid in the identification of sub-populations of individuals at risk for diabetes.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2022

Classification of bacterial plasmid and chromosome derived sequences using machine learning

Plasmids are important genetic elements that facilitate horizonal gene transfer between bacteria and contribute to the spread of virulence and antimicrobial resistance. Most bacterial genome sequences in the public archives exist in draft form with many contigs, making it difficult to determine if a contig is of chromosomal or plasmid origin. Using a training set of contigs comprising 10,584 chromosomes and 10,654 plasmids from the PATRIC database, we evaluated several machine learning models including random forest, logistic regression, XGBoost, and a neural network for their ability to classify chromosomal and plasmid sequences using nucleotide k-mers as features. Based on the methods tested, a neural network model that used nucleotide 6-mers as features that was trained on randomly selected chromosomal and plasmid subsequences 5kb in length achieved the best performance, outperforming existing out-of-the-box methods, with an average accuracy of 89.38% ± 2.16% over a 10-fold cross validation. The model accuracy can be improved to 92.08% by using a voting strategy when classifying holdout sequences. In both plasmids and chromosomes, subsequences encoding functions involved in horizontal gene transfer—including hypothetical proteins, transporters, phage, mobile elements, and CRISPR elements—were most likely to be misclassified by the model. This study provides a straightforward approach for identifying plasmid-encoding sequences in short read assemblies without the need for sequence alignment-based tools.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2023

An accurate and interpretable model for antimicrobial resistance in pathogenic Escherichia coli from livestock and companion animal species

Understanding the microbial genomic contributors to antimicrobial resistance (AMR) is essential for early detection of emerging AMR infections, a pressing global health threat in human and veterinary medicine. Here we used whole genome sequencing and antibiotic susceptibility test data from 980 disease causing Escherichia coli isolated from companion and farm animals to model AMR genotypes and phenotypes for 24 antibiotics. We determined the strength of genotype-to-phenotype relationships for 197 AMR genes with elastic net logistic regression. Model predictors were designed to evaluate different potential modes of AMR genotype translation into resistance phenotypes. Our results show a model that considers the presence of individual AMR genes and total number of AMR genes present from a set of genes known to confer resistance was able to accurately predict isolate resistance on average (mean F 1 score = 98.0%, SD = 2.3%, mean accuracy = 98.2%, SD = 2.7%). However, fitted models sometimes varied for antibiotics in the same class and for the same antibiotic across animal hosts, suggesting heterogeneity in the genetic determinants of AMR resistance. We conclude that an interpretable AMR prediction model can be used to accurately predict resistance phenotypes across multiple host species and reveal testable hypotheses about how the mechanism of resistance may vary across antibiotics within the same class and across animal hosts for the same antibiotic.

Chung, Henri C.↗
DOE OSTI2024

Misclassification of causes of death among a small all-autopsied group of former nuclear workers: Death certificates vs. autopsy reports

The U.S. Transuranium and Uranium Registries performs autopsies on each of its deceased Registrants as a part of its mission to follow up occupationally-exposed individuals. This provides a unique opportunity to explore death certificate misclassification errors, and the factors that influence them, among this small population of former nuclear workers. Underlying causes of death from death certificates and autopsy reports were coded using the 10 th revision of the International Classification of Diseases (ICD-10). These codes were then used to quantify misclassification rates among 268 individuals for whom both full autopsy reports and death certificates with legible underlying causes of death were available. When underlying causes of death were compared between death certificates and autopsy reports, death certificates correctly identified the underlying cause of death’s ICD-10 disease chapter in 74.6% of cases. The remaining 25.4% of misclassified cases resulted in over-classification rates that ranged from 1.2% for external causes of mortality to 12.2% for circulatory disease, and under-classification rates that ranged from 7.7% for external causes of mortality to 47.4% for respiratory disease. Neoplasms had generally lower misclassification rates with 4.3% over-classification and 13.3% under-classification. A logistic regression revealed that the odds of a match were 2.8 times higher when clinical history was mentioned on the autopsy report than when it was not. Similarly, the odds of a match were 3.4 times higher when death certificates were completed using autopsy findings than when autopsy findings were not used. This analysis excluded cases where it could not be determined if autopsy findings were used to complete death certificates. The findings of this study are useful to investigate the impact of death certificate misclassification errors on radiation risk estimates and, therefore, improve the reliability of epidemiological studies.

63 RADIATION, THERMAL, AND OTHER ENVIRON. POLLUTAN↗