Microbiome Data Accurately Predicts the Postmortem Interval Using Random Forest Regression Models
Aeriel Belk 1,†, Zhenjiang Zech Xu 2,†, David O. Carter 3, Aaron Lynne 4, Sibyl Bucheli 4
, Rob Knight 2,5,6,* and Jessica L. Metcalf 1,*
, Rob Knight 2,5,6,* and Jessica L. Metcalf 1,*
1
Department of Animal Sciences, Colorado State University, Fort Collins, CO 80525, USA
2
Department of Pediatrics, University of California, La Jolla, San Diego, CA 92093, USA
3
Laboratory of Forensic Taphonomy, Forensic Sciences Unit, Division of Natural Sciences and Mathematics, Chaminade University of Honolulu, Honolulu, HI 96816, USA
4
Department of Biological Sciences, Sam Houston State University, Huntsville, TX 77340, USA
5
Department of Computer Science and Engineering, University of California, San Diego, CA 92037, USA
6
Microbiome Innovation Center, University of California, San Diego, CA 92037, USA
*
Authors to whom correspondence should be addressed.
†
Contributed equally.
Genes 2018, 9(2), 104; https://doi.org/10.3390/genes9020104
Received: 31 December 2017 / Revised: 12 February 2018 / Accepted: 12 February 2018 / Published: 16 February 2018
(This article belongs to the Special Issue Forensic Genomics)
Abstract
Death investigations often include an effort to establish the postmortem interval (PMI) in cases in which the time of death is uncertain. The postmortem interval can lead to the identification of the deceased and the validation of witness statements and suspect alibis. Recent research has demonstrated that microbes provide an accurate clock that starts at death and relies on ecological change in the microbial communities that normally inhabit a body and its surrounding environment. Here, we explore how to build the most robust Random Forest regression models for prediction of PMI by testing models built on different sample types (gravesoil, skin of the torso, skin of the head), gene markers (16S ribosomal RNA (rRNA), 18S rRNA, internal transcribed spacer regions (ITS)), and taxonomic levels (sequence variants, species, genus, etc.). We also tested whether particular suites of indicator microbes were informative across different datasets. Generally, results indicate that the most accurate models for predicting PMI were built using gravesoil and skin data using the 16S rRNA genetic marker at the taxonomic level of phyla. Additionally, several phyla consistently contributed highly to model accuracy and may be candidate indicators of PMI. View Full-Text
Figure 1
This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited (CC BY 4.0).
Share & Cite This Article
MDPI and ACS Style
Belk, A.; Xu, Z.Z.; Carter, D.O.; Lynne, A.; Bucheli, S.; Knight, R.; Metcalf, J.L. Microbiome Data Accurately Predicts the Postmortem Interval Using Random Forest Regression Models. Genes 2018, 9, 104.
Note that from the first issue of 2016, MDPI journals use article numbers instead of page numbers. See further details here.