Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2023 Jan 20;11(2):98.
doi: 10.3390/toxics11020098.

A Machine Learning Model to Estimate Toxicokinetic Half-Lives of Per- and Polyfluoro-Alkyl Substances (PFAS) in Multiple Species

Daniel E Dawson  1 , Christopher Lau  2 , Prachi Pradeep  1   3 , Risa R Sayre  1 , Richard S Judson  1 , Rogelio Tornero-Velez  1 , John F Wambaugh  1
Affiliations

A Machine Learning Model to Estimate Toxicokinetic Half-Lives of Per- and Polyfluoro-Alkyl Substances (PFAS) in Multiple Species

Daniel E Dawson et al. Toxics. .

Abstract

Per- and polyfluoroalkyl substances (PFAS) are a diverse group of man-made chemicals that are commonly found in body tissues. The toxicokinetics of most PFAS are currently uncharacterized, but long half-lives (t½) have been observed in some cases. Knowledge of chemical-specific t½ is necessary for exposure reconstruction and extrapolation from toxicological studies. We used an ensemble machine learning method, random forest, to model the existing in vivo measured t½ across four species (human, monkey, rat, mouse) and eleven PFAS. Mechanistically motivated descriptors were examined, including two types of surrogates for renal transporters: (1) physiological descriptors, including kidney geometry, for renal transporter expression and (2) structural similarity of defluorinated PFAS to endogenous chemicals for transporter affinity. We developed a classification model for t½ (Bin 1: <12 h; Bin 2: <1 week; Bin 3: <2 months; Bin 4: >2 months). The model had an accuracy of 86.1% in contrast to 32.2% for a y-randomized null model. A total of 3890 compounds were within domain of the model, and t½ was predicted using the bin medians: 4.9 h, 2.2 days, 33 days, and 3.3 years. For human t½, 56% of PFAS were classified in Bin 4, 7% were classified in Bin 3, and 37% were classified in Bin 2. This model synthesizes the limited available data to allow tentative extrapolation and prioritization.

Keywords: PFAS; half-life; machine learning model; perfluoro-alkyl substances; toxicokinetics.

PubMed Disclaimer

Conflict of interest statement

The authors declare no conflict of interest.

Figures

Figure 1
Figure 1
Scientific workflow including (A) Training Data Assembly, (B) Predictor Dataset Assembly, (C) Dataset Processing and Preparation, (D) Random Forest Model Construction, and (E) Application of the Models to the CCD PFAS list. Green boxes denote data sources, purple boxes denote assembled datasets, red boxes denote models, blue boxes data denote processing steps, black boxes denote model outputs, and arrows indicate flow between steps.
Figure 2
Figure 2
Values of t½ of the training data (y-axis) vs. classification predictions by the RF Classification model using 15 predictors. Colors signify species, while shapes indicate different PFAS compounds. Bin margins (<12 h, 12 h−1 week, 1 week–2 months, >2 months) are indicated as dotted lines. Note that observations have been jittered (that is, a small amount of random variation has been added) along the x-axis to increase readability.
Figure 3
Figure 3
Distributions of predicted t½ for (A) 4136 PFAS within the AD of the RF Classification model, and (B) 921 PFAS classified in the same 3 classes as the 11 training set chemicals via ClassyFire. Shown are the number of chemicals predicted to fall within half-life categories by sex (male = ♂, female = ♀) for 5 species. Bins are denoted by color, with pink ≤ 12 h, green = 12 h−1 week, blue = 1 week−2 months, and purple ≥ 2 months.
See this image and copyright information in PMC

Similar articles

See all similar articles

Cited by

See all "Cited by" articles

References

    1. DeWitt J.C. Toxicological Effects of Perfluoroalkyl and Polyfluoroalkyl Substances. Springer; Berlin/Heidelberg, Germany: 2015.
    1. Buck R.C., Murphy P.M., Pabon M. Polyfluorinated Chemicals and Transformation Products. Springer; Berlin/Heidelberg, Germany: 2012. Chemistry, properties, and uses of commercial fluorinated surfactants; pp. 1–24.
    1. Rao N.S., Baker B.E. Organofluorine Chemistry. Springer; Berlin/Heidelberg, Germany: 1994. Textile finishes and fluorosurfactants; pp. 321–338.
    1. Rayne S., Forest K., Friesen K.J. Congener-specific numbering systems for the environmentally relevant C4 through C8 perfluorinated homologue groups of alkyl sulfonates, carboxylates, telomer alcohols, olefins, and acids, and their derivatives. J. Environ. Sci. Health Part A. 2008;43:1391–1401. doi: 10.1080/10934520802232030. - DOI - PubMed
    1. Wang Z., Cousins I.T., Scheringer M., Hungerbühler K. Fluorinated alternatives to long-chain perfluoroalkyl carboxylic acids (PFCAs), perfluoroalkane sulfonic acids (PFSAs) and their potential precursors. Environ. Int. 2013;60:242–248. doi: 10.1016/j.envint.2013.08.021. - DOI - PubMed

LinkOut - more resources