Comparing the potential health effects of chemicals

An optimized statistical tool allows for the accurate ranking of samples based on toxicity, according to researchers from the Division of Translational Toxicology.

The Toxicological Prioritization Index (ToxPi) is a statistical modeling framework originally developed for visual assessment of relative toxicity of multiple chemicals across an array of biological and physiochemical endpoints. For a given chemical space (the chemicals or substances being compared in a ToxPi analysis), ToxPi integrates the data into a set of weighted slices that provide a visual profile and a score metric for comparison, allowing for effective sample ranking among chemicals of interest. Traditionally, the weights on different measured endpoints have been user-supplied and not data-driven.

The researchers developed a methodology to optimize ToxPi model weights with user input on the relative bioactivity or toxicity of some of the chemicals in the experiment. This input is then used with ordinal regression and a custom genetic algorithm to compute optimal model weights. Ordinal regression provides an ordered classification of samples, and genetic algorithms are used to find the best solution from all feasible solutions based on natural selection properties. This allows for improved classification and ranking of the chemicals.

A simulation study revealed that the genetic algorithm can greatly improve the ranking results of ordinal regression for less complex models with smaller slice numbers. The researchers also built a high-performance ToxPi model that can predict petroleum substance bioactivity using available polycyclic aromatic compound data. Overall, the results showed that ordinal regression and the custom genetic algorithm were accurate methods for predicting ToxPi rankings across an array of common use case scenarios. (JW)

Citation: Fleming JF, House JS ,Chappel JR, Motsinger-Reif AA, DM Reif DM. 2024. Guided optimization of ToxPi model weights using a semi-automated approach. Comput Toxicol 29:100294.

Leisure-time physical activity reduces breast cancer risk in young women

Engaging in leisure-time physical activity may help to reduce the risk of breast cancer in premenopausal women, according to NIEHS researchers and their collaborators from the Institute of Cancer Research in London. Leisure-time physical activity is defined as any activity not required as part of everyday life, such as sports, exercise, or recreational walking.

Breast cancer is currently the most commonly diagnosed form of cancer among reproductive-age women worldwide. Although leisure-time physical activity can reduce risk of postmenopausal breast cancer, its effect on premenopausal breast cancer is less clear.

To address this knowledge gap, the researchers conducted a pooled analysis of 19 cohort studies in the Premenopausal Breast Cancer Collaborative Group. The analysis included 547,601 women, including 10,231 who developed breast cancer. The authors estimated the relative risk of breast cancer among women in the “most active” group (90th percentile) compared to the “least active” group (10th percentile). After accounting for body mass index and other factors, the most active were 10% less likely to develop breast cancer before menopause. For HER2-enriched breast cancer, the risk reduction was 43%.

Overall, there was a consistent trend of lower breast cancer rates being associated with increasing levels of physical activity. According to the authors, the results add to evidence that engagement in higher levels of leisure-time physical activity may lead to reduced premenopausal breast cancer risk. [Read related article.] (BR)

Citation: Timmins IR, Jones ME, O’Brien KM, HO Adami, Aune D, Baglietto L, Bertrand KA, Brantley KD, Chen Y, DeHart JC, Clendenen TV, Dossus L, Eliassen AH, Fletcher O, Fournier A, Hakansson N, Hankinson SE, Houlston RS, Joshu CE, Kirsh, Kitahara CM, Koh WP, Linet MS, Park HL, Lynch BM, May AM, Mellemkjær L, Milne RL, Palmer JR, Ricceri F, Rohan TE, Ruddy KJ, Sanchez MJ, Shu XO, Smith-Byrne K, Steindorf K, Sund M, Vachon CM, Vatten LJ, Visvanathan K, Weiderpass E, Willet WC, Wolk A, Yuan JM, Wei Z, Nichols HB, Sandler DP, Swerdlow AJ, Schoemaker MJ. 2023. International pooled analysis of leisure-time physical activity and premenopausal breast cancer in women from 19 cohorts. J Clin Oncol 23:JCO2301101.

Simplifying complex biological datasets

A novel computational approach called MIBCOVIS could enable deeper insights into complicated biological systems, according to NIEHS researchers and their collaborators.

Many types of complex datasets characterized by intricate patterns and variations that change over time or space are difficult to simplify, visualize, and interpret. For example, in single-cell analyses used in scientific fields such as developmental biology, the true underlying patterns are often unknown or hard to measure. This lack of a “ground truth” makes it challenging to directly evaluate accuracy, hindering interpretations of complicated biological interactions.

To address the need for models and algorithms capable of analyzing complicated datasets, the researchers developed MIBCOVIS as a benchmarking tool to optimally visualize and interpret high-dimensional data lacking ground truth. MIBCOVIS enables easy comparisons of the performance of various data reduction methods. Leveraging MIBCOVIS, the researchers evaluated seven methods applied to four biological processes. They found that these existing methods do not simultaneously optimize the visualization and interpretability of the datasets. According to the researchers, this powerful framework can be applied to single-cell and non–single-cell data from various scientific fields. (JW)

Citation: Atitey K, Motsinger-Reif AA, Anchang B. 2023. Model-based evaluation of spatiotemporal data reduction methods with unknown ground truth through optimal visualization and interpretability metrics. Brief Bioinform 25(1):bbad455.

Majority of clinicians do not frequently assess environmental asthma triggers

Environmental assessment and recommendations to patients vary considerably among asthma care providers, according to NIEHS researchers and their collaborators.

Control of environmental factors is an important part of asthma management and can reduce asthma-related symptoms and mortality. Asthma guidelines recommend avoidance and reduction of exposures that exacerbate the disease, including exposures to allergens, tobacco smoke, air pollution, and other irritants. Yet limited national-level data exist on how clinicians implement environmental control recommendations.

To address this knowledge gap, the researchers analyzed data on clinicians’ self-reported use of recommended environmental control practices in a nationally representative sample of 1,645 primary care physicians, asthma specialists, and advanced practice providers. The results revealed that a higher percentage of specialists assessed asthma triggers at home, school, or work than primary care or advanced practice providers.

However, 46%-76% of clinicians, depending on clinician type, reported not assessing triggers almost always during asthma visits. Nearly all clinicians recommended avoiding secondhand tobacco smoke, and most clinicians reported recommending pet removal, especially from homes of patients with pet sensitivities. In contrast, advice regarding cooking appliances, including proper ventilation, was infrequent. According to the authors, the results helped to identify areas where guideline uptake was suboptimal. The findings suggest that group-specific studies might inform targeted guideline implementation interventions and improve patient care in the future. (JW)

Citation: Salo PM, Akinbami LJ, Cloutier MM, Wilkerson JC, Elward KS, Mazurek JM, Diette GB, Mitchell TA, Williams S, Zeldin DC. 2023. Environmental management of asthma in clinical practice: results from the 2012 National Ambulatory Medical Care Survey. J Allergy Clin Immunol Glob 3(1):100192.

Phthalates linked to increased rates of preterm birth among marginalized groups

Exposure to chemicals called phthalates during pregnancy may contribute to racial and ethnic disparities in preterm birth, according to NIEHS researchers and their collaborators.

Prenatal exposure to synthetic chemicals, including phthalates, is increasingly considered an important risk factor for preterm birth. Phthalates are used extensively in commercial goods, such as personal care products, food packaging materials, and medications. Because of their widespread use, urinary phthalate metabolites are ubiquitous in the U.S. population, and exposure may be especially concerning among pregnant individuals. There are also recognized racial and ethnic disparities in phthalate exposure. Yet previous studies that characterized racial and ethnic differences in phthalate exposure during pregnancy in the U.S. have been limited by relatively small sample sizes, and none have rigorously investigated how exposure disparities could contribute to the concurrent racial and ethnic disparities in rates of preterm birth.

To overcome this limitation, the researchers pooled individual-level data on 6,045 pregnancies from 16 U.S. cohorts and investigated differences in nine urinary phthalate metabolite concentrations by race and ethnicity. Compared with White participants, mean phthalate metabolite concentrations were consistently higher among Black and Hispanic/Latina participants by as much as 148% and 94%, respectively. The probability of preterm birth among these groups would be approximately 13% and 9% lower, respectively, if they had urinary phthalate metabolite concentrations approximately equal to those among White participants. According to the authors, these findings support the need to reduce phthalate exposures among systematically marginalized people. (JW)

Citation: Welch BM, Keil AP, Buckley JP, Engel SM, James-Todd T, Zota AR, Alshawabkeh AN, Barrett ES, Bloom MS, Bush NR, Cordero JF, Dabelea D, Eskenazi B, Lanphear BP, Padmanabhan V, Sathyanarayana S, Swan SH, Aalborg J, Baird DD, Binder AM, Bradman A, Braun JM, Calafat AM, Cantonwine DE, Christenbury KE, Factor-Litvak P, Harley KG, Hauser R, Herbstman JB, Hertz-Picciotto I, Holland N, Jukic AMZ, McElrath TF, Meeker JD, Messerlian C, Michels KB, Newman RB, Nguyen RHN, O'Brien KM, Rauh VA, Redmon B, Rich DQ, Rosen EM, Schmidt RJ, Sparks AE, Starling AP, Wang C, Watkins DJ, Weinberg CR, Weinberger B, Wenzel AG, Wilcox AJ, Yolton K, Zhang Y, Ferguson KK. 2023. Racial and ethnic disparities in phthalate exposure and preterm birth: a pooled study of sixteen U.S. cohorts. Environ Health Perspect 131(12):127015.