Skip to main content

A preliminary study on early postmortem submersion interval (PMSI) estimation and cause-of-death discrimination based on nontargeted metabolomics and machine learning algorithms


Postmortem submersion interval (PMSI) estimation and cause-of-death discrimination of corpses in water have long been challenges in forensic practice. Recently, many studies have linked postmortem metabolic changes with PMI extension, providing a potential strategy for estimating PMSI using the metabolome. Additionally, there is a lack of potential indicators with high sensitivity and specificity for drowning identification. In the present study, we profiled the untargeted metabolome of blood samples from drowning and postmortem submersion rats at different PMSIs within 24 h by liquid chromatography–tandem mass spectrometry (LC–MS/MS). A total of 601 metabolites were detected. Four different machine learning algorithms, including random forest (RF), partial least squares (PLS), support vector machine (SVM), and neural network (NN), were used to compare the efficiency of the machine learning methods. Nineteen metabolites with obvious temporal regularity were selected as candidate biomarkers according to “IncNodePurity.” Robust models were built with these biomarkers, which yielded a mean absolute error of 1.067 h. Additionally, 36 other metabolites were identified to build the classifier model for discriminating drowning and postmortem submersion (AUC = 1, accuracy = 95%). Our results demonstrated the potential application of metabolomics combined with machine learning in PMSI estimation and cause-of-death discrimination.

This is a preview of subscription content, access via your institution.

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5

Availability of data and material


Code availability

Not applicable.


  1. van Beeck EF, Branche CM, Szpilman D, Modell JH, Bierens JJ (2005) A new definition of drowning: towards documentation and prevention of a global public health problem. Bull WHO 83(11):853–856. doi:/S0042-96862005001100015

  2. World Health Organization (2021) Drowning. Accessed 27 May 2021

  3. Humphreys MK, Panacek E, Green W, Albers E (2013) Comparison of protocols for measuring and calculating postmortem submersion intervals for human analogs in fresh water. J Forensic Sci 58(2):513–517.

    Article  PubMed  Google Scholar 

  4. Heaton V, Lagden A, Moffatt C, Simmons T (2010) Predicting the postmortem submersion interval for human remains recovered from U.K. waterways. J Forensic Sci 55(2):302–307

  5. van Daalen MA, de Kat DS, Oude Grotebevelsborg BF, de Leeuwe R, Warnaar J, Oostra RJ, WL MD-H (2017) An aquatic decomposition scoring method to potentially predict the postmortem submersion interval of bodies recovered from the North Sea. J Forensic Sci 62(2):369–373

  6. Palazzo C, Pelletti G, Fais P, Boscolo-Berto R, Fersini F, Gaudio RM, Pelotti S (2020) Postmortem submersion interval in human bodies recovered from fresh water in an area of Mediterranean climate. Application and comparison of preexisting models. Forensic Sci Int 306:110051

  7. Oshima T, Ohtani M, Mimasaka S (2019) Muscular hemorrhages around the scapula resulting from excessive upper extremity motion in cases of fatal drowning: autopsy findings for insights on manner of death. Forensic Sci Int 300:82–84.

    Article  PubMed  Google Scholar 

  8. Perez-Carceles MD, del Pozo S, Sibon A, Noguera JA, Osuna E, Vizcaya MA, Luna A (2012) Serum biochemical markers in drowning: diagnostic efficacy of Strontium and other trace elements. Forensic Sci Int 214(1–3):159–166.

    CAS  Article  PubMed  Google Scholar 

  9. Di Giancamillo A, Giudici E, Andreola S, Porta D, Gibelli D, Domeneghini C, Grandi M, Cattaneo C (2010) Immersion of piglet carcasses in water–the applicability of microscopic analysis and limits of diatom testing on an animal model. Leg Med (Tokyo) 12(1):13–18.

    Article  Google Scholar 

  10. Locci E, Stocchero M, Noto A, Chighine A, Natali L, Napoli PE, Caria R, De-Giorgio F, Nioi M, d’Aloja E (2019) A (1)H NMR metabolomic approach for the estimation of the time since death using aqueous humour: an animal model. Metabolomics 15(5):76.

    CAS  Article  PubMed  Google Scholar 

  11. Brooks JW (2016) Postmortem changes in animal carcasses and estimation of the postmortem interval. Vet Pathol 53(5):929–940.

    CAS  Article  PubMed  Google Scholar 

  12. Locci E, Stocchero M, Gottardo R, De-Giorgio F, Demontis R, Nioi M, Chighine A, Tagliaro F, d’Aloja E (2021) Comparative use of aqueous humour (1)H NMR metabolomics and potassium concentration for PMI estimation in an animal model. Int J Legal Med 135(3):845–852.

    Article  PubMed  Google Scholar 

  13. Rosa MF, Scano P, Noto A, Nioi M, Sanna R, Paribello F, De-Giorgio F, Locci E, d’Aloja E (2015) Monitoring the modifications of the vitreous humor metabolite profile after death: an animal model. Biomed Res Int 2015:627201.

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  14. Nicholson JK, Lindon JC (2008) Systems biology: metabonomics. Nature 455(7216):1054–1056.

    CAS  Article  PubMed  Google Scholar 

  15. Kaszynski RH, Nishiumi S, Azuma T, Yoshida M, Kondo T, Takahashi M, Asano M, Ueno Y (2016) Postmortem interval estimation: a novel approach utilizing gas chromatography/mass spectrometry-based biochemical profiling. Anal Bioanal Chem 408(12):3103–3112.

    CAS  Article  PubMed  Google Scholar 

  16. Sato T, Zaitsu K, Tsuboi K, Nomura M, Kusano M, Shima N, Abe S, Ishii A, Tsuchihashi H, Suzuki K (2015) A preliminary study on postmortem interval estimation of suffocated rats by GC-MS/MS-based plasma metabolic profiling. Anal Bioanal Chem 407(13):3659–3665.

    CAS  Article  PubMed  Google Scholar 

  17. Bierens JJ, Lunetta P, Tipton M, Warner DS (2016) Physiology of drowning: a review. Physiology (Bethesda) 31(2):147–166.

    Article  Google Scholar 

  18. Bohnert S, Reinert C, Trella S, Schmitz W, Ondruschka B, Bohnert M (2021) Metabolomics in postmortem cerebrospinal fluid diagnostics: a state-of-the-art method to interpret central nervous system-related pathological processes. Int J Legal Med 135(1):183–191.

    Article  PubMed  Google Scholar 

  19. Arends D, van der Velde KJ, Prins P, Broman KW, Moller S, Jansen RC, Swertz MA (2012) xQTL workbench: a scalable web environment for multi-level QTL analysis. Bioinformatics 28(7):1042–1044.

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  20. Liebal UW, Phan ANT, Sudhakar M, Raman K, Blank LM (2020) Machine learning applications for mass spectrometry-based metabolomics. Metabolites 10(6)

  21. He F, Huang D, Liu L, Shu X, Yin H, Li X (2008) A novel PCR-DGGE-based method for identifying plankton 16S rDNA for the diagnosis of drowning. Forensic Sci Int 176(2–3):152–156.

    CAS  Article  PubMed  Google Scholar 

  22. Dunn WB, Wilson ID, Nicholls AW, Broadhurst D (2012) The importance of experimental design and QC samples in large-scale and MS-driven untargeted metabolomic studies of humans. Bioanalysis 4(18):2249–2264.

    CAS  Article  PubMed  Google Scholar 

  23. Han C, Jiang YH, Li W, Liu Y, Qi ZQ (2019) Study on the antihypertensive mechanism of Astragalus membranaceus and Salvia miltiorrhiza based on intestinal flora-host metabolism. Evid Based Complement Alternat Med 2019:5418796.

    Article  PubMed  PubMed Central  Google Scholar 

  24. Breiman L (2001) Random Forests. Machine Learning 45(1):5–32.

    Article  Google Scholar 

  25. Pan L, Liu G, Lin F, Zhong S, Xia H, Sun X, Liang H (2017) Machine learning applications for prediction of relapse in childhood acute lymphoblastic leukemia. Sci Rep 7(1):7402.

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  26. Caruana R, Niculescu-Mizil A (2006) An empirical comparison of supervised learning algorithms. In: Proceedings of the 23rd International Conference on Machine Learning. The Twenty-third International Conference on Machine Learning,  Pittsburgh Pennsylvania USA, June 25-29 2006, pp 161–168.

  27. Liaw A, Wiener M (2002) Classification and regression by random. For R News 2(3):18–22

    Google Scholar 

  28. Castiglione F, Ferro M, Mavroudakis E, Pellitteri R, Bossolasco P, Zaccheo D, Morbidelli M, Silani V, Mele A, Moscatelli D, Cova L (2017) NMR metabolomics for stem cell type discrimination. Sci Rep 7(1):15808.

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  29. Swain R, Kumar A, Sahoo J, Lakshmy R, Gupta SK, Bhardwaj DN, Pandey RM (2015) Estimation of post-mortem interval: a comparison between cerebrospinal fluid and vitreous humour chemistry. J Forensic Leg Med 36:144–148.

    Article  PubMed  Google Scholar 

  30. Dai X, Fan F, Ye Y, Lu X, Chen F, Wu Z, Liao L (2019) An experimental study on investigating the postmortem interval in dichlorvos poisoned rats by GC/MS-based metabolomics. Leg Med (Tokyo) 36:28–36.

    CAS  Article  Google Scholar 

  31. Zhang J, Li B, Wang Q, Wei X, Feng W, Chen Y, Huang P, Wang Z (2017) Application of Fourier transform infrared spectroscopy with chemometrics on postmortem interval estimation based on pericardial fluids. Sci Rep 7(1):18013.

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  32. Jorkasky DK (2010) Biomarkers in drug development: a handbook of practice, application and strategy. Br J Clin Pharmacol 70(1):151–152.

    Article  PubMed Central  Google Scholar 

  33. Zelentsova EA, Yanshole LV, Melnikov AD, Kudryavtsev IS, Novoselov VP, Tsentalovich YP (2020) Post-mortem changes in metabolomic profiles of human serum, aqueous humor and vitreous humor. Metabolomics 16(7):80.

    CAS  Article  PubMed  Google Scholar 

  34. Yu Q, Tian X, Shao L, Li X, Dai R (2019) Targeted metabolomics to reveal muscle-specific energy metabolism between bovine longissimus lumborum and psoas major during early postmortem periods. Meat Sci 156:166–173.

    CAS  Article  PubMed  Google Scholar 

  35. Zelentsova EA, Yanshole LV, Snytnikova OA, Yanshole VV, Tsentalovich YP, Sagdeev RZ (2016) Post-mortem changes in the metabolomic compositions of rabbit blood, aqueous and vitreous humors. Metabolomics 12(11):172.

    CAS  Article  Google Scholar 

  36. Jawor P, Zabek A, Wojtowicz W, Krol D, Stefaniak T, Mlynarz P (2019) Metabolomic studies as a tool for determining the post-mortem interval (PMI) in stillborn calves. BMC Vet Res 15(1):189.

    Article  PubMed  PubMed Central  Google Scholar 

  37. Donaldson AE, Lamont IL (2015) Metabolomics of post-mortem blood: identifying potential markers of post-mortem interval. Metabolomics 11(1):237–245.

    CAS  Article  Google Scholar 

  38. Wishart DS, Jewison T, Guo AC, Wilson M, Knox C, Liu Y, Djoumbou Y, Mandal R, Aziat F, Dong E, Bouatra S, Sinelnikov I, Arndt D, Xia J, Liu P, Yallou F, Bjorndahl T, Perez-Pineiro R, Eisner R, Allen F, Neveu V, Greiner R, Scalbert A (2013) HMDB 3.0--The human metabolome database in 2013. Nucleic Acids Res 41(Database issue):D801–D807

  39. Pesko BK, Weidt S, McLaughlin M, Wescott DJ, Torrance H, Burgess K, Burchmore R (2020) Postmortomics: the potential of untargeted metabolomics to highlight markers for time since death. OMICS 24(11):649–659.

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  40. Skeaff S, Te Morenga L (2017) Nutrition Society of New Zealand Annual Conference Held in Wellington, New Zealand, 1-4 December 2015. Nutrients 9(3)

  41. Sinclair GM, O'Brien AL, Keough M, de Souza DP, Dayalan S, Kanojia K, Kouremenos K, Tull DL, Coleman RA, Jones OAH, Long SM (2019) Metabolite changes in an estuarine annelid following sublethal exposure to a mixture of zinc and boscalid. Metabolites 9(10)

  42. Daly R, Blackburn G, Best C, Goodyear CS, Mudaliar M, Burgess K, Stirling A, Porter D, McInnes IB, Barrett MP, Dale J (2020) Changes in plasma itaconate elevation in early rheumatoid arthritis patients elucidates disease activity associated macrophage activation. Metabolites 10(6)

  43. Ustunel I, Acar N, Gemici B, Ozbey O, Edizer I, Soylu H, Tepekoy F, Izgut-Uysal VN (2014) The effects of water immersion and restraint stress on the expressions of apelin, apelin receptor (APJR) and apoptosis rate in the rat heart. Acta Histochem 116(5):675–681.

    CAS  Article  PubMed  Google Scholar 

  44. Hayward JS, Eckerson JD (1984) Physiological responses and survival time prediction for humans in ice-water. Aviat Space Environ Med 55(3):206–211

    CAS  PubMed  Google Scholar 

  45. Piette MH, De Letter EA (2006) Drowning: still a difficult autopsy diagnosis. Forensic Sci Int 163(1–2):1–9.

    Article  PubMed  Google Scholar 

  46. Wang LL, Zhang FY, Dong WW, Wang CL, Liang XY, Suo LL, Jian C, Zhang M, Guo XS, Jiang PH, Guan DW, Zhao R (2020) A novel approach for the forensic diagnosis of drowning by microbiological analysis with next-generation sequencing and unweighted UniFrac-based PCoA. Int J Legal Med 134(6):2149–2159.

    Article  PubMed  Google Scholar 

  47. Du T, Lin Z, Xie Y, Ye X, Tu C, Jin K, Xie J, Shen Y (2018) Metabolic profiling of femoral muscle from rats at different periods of time after death. PLoS One 13(9):e0203920.

    CAS  Article  PubMed  PubMed Central  Google Scholar 

Download references


This study was financially supported by the National Key Research and Development Program of China (Grant No. 2018YFC0807204) and National Natural Science Foundation of China (Grant Nos. 81871529, 81971793, 81801874, 81671862).

Author information

Authors and Affiliations



D Guan and R Zhao conceived and designed the research. F Zhang and L Wang performed the lab experiments and wrote the main manuscript text. W Dong, M Zhang, D Tash, and X Li performed the animal experiments. F Zhang, S Du, and H Yuan performed the bioinformatic analysis. All authors have read and commented on the manuscript.

Corresponding authors

Correspondence to Rui Zhao or Da-Wei Guan.

Ethics declarations

Ethics approval

The use of animals was approved by the Animal Experiment Committee of China Medical University. All experiments were conducted according to the guidelines of the National Institutes of Health Guide for the Care and Use of Laboratory Animals.

Conflict of interest

The authors declare no competing interests.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Fu-Yuan Zhang and Lin-Lin Wang are contributed equally

Supplementary Information

Below is the link to the electronic supplementary material.

Supplementary file1 (DOCX 2949 KB)

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Zhang, FY., Wang, LL., Dong, WW. et al. A preliminary study on early postmortem submersion interval (PMSI) estimation and cause-of-death discrimination based on nontargeted metabolomics and machine learning algorithms. Int J Legal Med 136, 941–954 (2022).

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI:


  • Forensic medicine
  • Metabolomics
  • Drowning
  • Postmortem submersion interval
  • Machine learning