Evaluating Bioinformatic Pipeline Performance for Forensic Microbiome Analysis*,†,‡

Microbial communities have potential evidential utility for forensic applications. However, bioinformatic analysis of high-throughput sequencing data varies widely among laboratories. These differences can potentially affect microbial community composition and downstream analyses. To illustrate the importance of standardizing methodology, we compared analyses of postmortem microbiome samples using several bioinformatic pipelines, varying minimum library size or minimum number of sequences per sample, and sample size. Using the same input sequence data, we found that three open-source bioinformatic pipelines, MG-RAST, mothur, and QIIME2, had significant differences in relative abundance, alpha-diversity, and beta-diversity, despite the same input data. Increasing minimum library size and sample size increased the number of low-abundant and infrequent taxa detected. Our results show that bioinformatic pipeline and parameter choice affect results in important ways. Given the growing potential application of forensic microbiology to the criminal justice system, continued research on standardizing computational methodology will be important for downstream applications.     

A survey of bacteria associated with various life stages of primary colonizers: Lucilia sericata and Phormia regina

Blow flies are common primary colonizers of carrion, play an important role in the transfer of microbes between environments, and serve as a vector for many human pathogens. While some investigation has begun regarding the bacteria associated with different life stages of blow flies, a well replicated study is currently not available for the majority of blow flies. This study investigated bacteria associated with successive life stages of blow fly species Lucilia sericata and Phormia regina. A total of 38 samples were collected from four true replicates of L. sericata and P. regina. Variable region four (V4) of 16S ribosomal DNA (16S rDNA) was amplified and sequenced on MiSeq FGx sequencing platform using universal 16S rDNA primers and dual-index sequencing strategy. Bacterial communities associated with different life stages of L. sericata and P. regina didn't differ significantly from each other. In both blow fly species, Bacilli (e.g., Lactococcus) and Gammaproteobacteria (e.g., Providencia) constituted >95% of all bacterial classes across all life stages. At the genus level, Vagococcus and Leuconostoc were present at relatively high abundances in L. sericata whereas Yersinia and Proteus were present at comparatively high abundances in P. regina. Overall, information on bacterial structures associated with various life stages of blow flies can help scientists in better understanding or management of vector-borne pathogen dispersal and in increasing the accuracy of microbial evidence based postmortem interval (PMI) prediction models.     

Predicting the postmortem interval using human intestinal microbiome data and random forest algorithm

Gradual changes in microbial communities in a human body after death can be used to determine postmortem interval (PMI). In this study, gut microflora samples were collected from the vermiform appendix and the transverse colon of human cadavers with PMIs between 5 and 192 h. The results revealed that the appendix might be an excellent intestinal sampling site and the appendix flora had an inferred succession rule during human body decomposition. Firmicutes, Bacteroidetes, and their respective subclasses showed a predictable succession rule in relative abundance over time. A Random Forest regression model was developed to correlate human gut microbiota with PMI. We believe that our findings have increased the knowledge of the composition and abundance of the gut microbiota in human corpses, and suggest that the use of the human appendix microbial succession may be a potential method for forensic estimation of the time of death.     

Identifying background microbiomes in an evidence recovery laboratory: A preliminary study

16S rRNA profiling of bacterial communities may have forensic utility in the identification or association of individuals involved with criminal activities. Microbial profiling of evidence may, in the future, be performed within environments currently utilised for human DNA recovery, such as a forensic biology laboratory. It would be important to establish the background microbiome of such an environment to determine the potential presence of human or environmental microbial signatures to assist forensic scientists in the appropriate interpretation of target microbial communities. This study sampled various surfaces of an Evidence Recovery Laboratory (ERL) on three occasions including (a) before a monthly deep-clean, (b) immediately following the deep-clean, and (c) immediately after the laboratory’s use by a single participant for the purposes of routine item examinations. Microbial profiles were also generated for the involved participant and researcher for comparison purposes. Additionally, human nuclear DNA was profiled for each of the samples collected, using standard forensic profiling techniques, to provide a prospective link to the presence or absence of a background microbial signature within the ERL after its use. Taxonomic distributions across ERL samples revealed no consistent signature of any of the items sampled over time, however, major phyla noted within all ERL samples across the three timepoints were consistent with those found in human skin microbiomes. PCoA plots based on the Unweighted Unifrac metric revealed some clustering between participant microbial reference samples and surfaces of the ERL after use, suggesting that despite a lack of direct contact, and adherence to standard operating procedures (SOPs) suitable for human DNA recovery, microbiomes may be deposited into a forensic setting over time. The reference samples collected from the involved participant and researcher generated full STR profiles. Human DNA was observed to varying degrees in samples taken from the ERL across each of the sampling timepoints. There was no correlation observed between samples that contained or did not contain detectable quantities of human nuclear DNA and microbial profile outputs.     

Ensuring the safe and effective FDA regulation of fecal microbiota transplantation

Scientists, policymakers, and medical professionals alike have become increasingly worried about the rise of antibiotic resistance, and the growing number of infections due to bacteria like Clostridium difficile, which cause a significant number of deaths and are imposing increasing costs on our health care system. However, in the last few years, fecal microbiota transplantation (FMT), the transplantation of stool from a healthy donor into the bowel of a patient, has emerged as a startlingly effective means to treat recurrent C. difficile infections. At present, the FDA is proposing to regulate FMT as a biologic drug. However, this proposed classification is both underregulatory and overregulatory. The FDA''s primary goal is to ensure that patients have access to safe, effective treatments—and as such they should regulate some aspects of FMT more stringently than they propose to, and others less so. This essay will examine the nature of the regulatory challenges the FDA will face in deciding to regulate FMT as a biologic drug, and will then evaluate available policy alternatives for the FDA to pursue, ultimately concluding that the FDA ought to consider adopting a hybrid regulatory model as it has done in the case of cord blood.     

High resolution melting analysis (HRM) based on 16SrRNA as a tool for personal identification with the human oral microbiome

Personal identification plays an important role in the forensic practice. The conventional methods of personal identification including STR, SNP and InDel are focusing on the molecular characteristics of the human cell. Recently, researchers pay attention to human microbiome by the reason that the human microbiome is rich in amount and variable among people. The purpose of this study is to apply the human oral microbiome to forensic personal identification. We designed one general primer pair: 518 F&806R, which targeted the 16SrRNA. We conducted the high-resolution melting analysis (HRM) with the one primer pairs. The results indicated that oral microbiome from different people could be distinguished by using HRM based on 16SrRNA. This study showed that human oral microbiome could be a promising marker for the forensic personal identification application.     

Trace Evidence Potential in Postmortem Skin Microbiomes: From Death Scene to Morgue

Microbes can be used effectively as trace evidence, at least in research settings. However, it is unknown whether skin microbiomes change prior to autopsy and, if so, whether these changes interfere with linking objects to decedents. The current study included microbiomes from 16 scenes of death in the City and County of Honolulu and tested whether objects at the scenes can be linked to individual decedents. Postmortem skin microbiomes were stable during repeated sampling up to 60 h postmortem and were similar to microbiomes of an antemortem population. Objects could be traced to decedents approximately 75% of the time, with smoking pipes and medical devices being especially accurate (100% match), house and car keys being poor (0%), and other objects like phones intermediate (~80%). These results show that microbes from objects at death scenes can be matched to individual decedents, opening up a new method of establishing associations and identifications.     

Comparison of Commercial Kits for Recovery and Analysis of Bacterial DNA From Fingerprints

In forensic science, fingerprints are a common source of evidentiary information. However, latent examination is not always successful and trace human DNA cannot always be obtained. Thus, examining the fingerprint microbiome may offer a suitable alternative to more traditional methods of forensic identification. The Zymo Research ZR Bacterial/Fungal DNA MicroPrep™ Kit, Qiagen QIAmp^® DNA Mini Kit, Promega Wizard^® Genomic DNA Purification Kit, and the MPBio FastDNA^® Spin Kit were compared for their ability to yield a sufficient amount of bacterial DNA for next-generation sequencing in order to obtain a microbiome profile. Prints were deposited onto slides, allowed to sit for up to 1 month, and total DNA isolated and quantified using each kit. The kit from Zymo Research yielded the most concentrated DNA sample (0.0084 ng/µL) in the least amount of time as compared to other kits examined. Although this amount of DNA was far below the recommended DNA concentration threshold recommended for next-generation sequencing, a microbiome profile was successfully obtained. As interest in using the microbiome of an individual as a forensic tool continues to increase, there is the possibility that the microbiome of a fingerprint could complement traditional human DNA profiling in the future. This study provides evidence that trace amounts of bacterial DNA from fingerprints is quantifiable and sufficient for microbiome analysis.     

Estimating Time Since Death from Postmortem Human Gut Microbial Communities

Postmortem succession of human‐associated microbial communities (“human microbiome”) has been suggested as a possible method for estimating postmortem interval (PMI) for forensic analyses. Here we evaluate human gut bacterial populations to determine quantifiable, time‐dependent changes postmortem. Gut microflora were repeatedly sampled from the proximal large intestine of 12 deceased human individuals as they decayed under environmental conditions. Three intestinal bacterial genera were quantified by quantitative PCR (qPCR) using group‐specific primers targeting 16S rRNA genes. Bacteroides and Lactobacillus relative abundances declined exponentially with increasing PMI at rates of N_t = 0.977e^?0.0144t (r² = 0.537, p < 0.001) and N_t = 0.019e^?0.0087t (r² = 0.396, p < 0.001), respectively, where N_t is relative abundance at time (t) in cumulative degree hours. Bifidobacterium relative abundances did not change significantly: N_t = 0.003e^?0.002t (r² = 0.033, p = 0.284). Therefore, Bacteroides and Lactobacillus abundances could be used as quantitative indicators of PMI.