Using machine learning to predict soil lead relative bioavailability

J Hazard Mater. 2024 Nov 22:483:136515. doi: 10.1016/j.jhazmat.2024.136515. Online ahead of print.

Abstract

Although the relative bioavailability (RBA) can be applied to assess the effects of Pb on human health, there is no definition and no specific data of Pb-RBA to different soil sources and endpoints in vivo. In this study, we estimated the Pb-RBA from different soil sources and endpoints based on machine learning. The Pb-BAc and Pb-RBA in soils were found to be mostly in the range of 20-80 %, which is different from the USEPA Pb-RBA of 60 % in soils. The mean Pb-RBA for different biological endpoints in vivo predicted using the RF model were 49.94 ± 18.65 % for blood; 60.15 ± 26.62 %, kidney; 60.90 ± 21.51 %, liver; 50.70 ± 17.56 %, femur; and 62.89 ± 16.64 % as a combined measure. Pb-RBA of shooting range soils was 88.21 ± 16.92 % (mean), spiked/aged soils 77.11 ± 14.05 % and certified reference materials 73.70 ± 20.31 %; agricultural soil 68.28 ± 18.93 %, urban soil 64.36 ± 21.82 %, mining/smelting soils 53.99 ± 17.66 %, and industrial soils 47.71 ± 20.35 %. This study is first to define the Pb-RBA according to various soil sources and endpoints in vivo with the objective of providing more accurate Pb-RBA data for soil lead risk assessment.

Keywords: Bioaccessibility (Pb-BAc); In vivo-in vitro correlations (IVIVCs); Machine learning; Pb; Relative bioavailability (Pb-RBA).