7/25/10 Soil Forensics Soil components •  Abio/c parent material: stable over /me, affected slightly by climate, weather (slow processes) –  Can look at elemental analyses, organic ma?er, rare elements/glass/rubber, etc. •  Biological frac/on: flora and fauna –  Metagenomic DNA •  How can these components be used for forensics? •  SOIL IS THE ULTIMATE MIXTURE SAMPLE!!! Abio/c analyses •  Forensic geologists: X-­‐ray diffrac/on (mineral content), Infrared spectroscopy (organic frac/on); ICP-­‐MS (elemental composi/on) •  Fatal car crash, suspects flee down river bank; apprehended hours later; deny being in the area mud on shoe •  Sample taken from shoe print at river bank, Munsell color chart, microscopic/morphological comparisons = put suspect at crime scene even though he was denying ever being near the river 1 7/25/10 Forensic comparison of soils, Fitzpatrick, et al. 2009 Next steps •  Sulfur par9cles and content similar (Munsell) •  Xray Diffrac9on (XRD) paDerns similar •  Diffuse Reflectance Infrared Fourier Transform spectroscopy (DRIFT) –  collects and analyzes scaDered IR energy Compared to alibi loca/on soil •  The shoe and shoe print had higher similarity across all measures than they did with the alibi soil •  Suspect found guilty of hit and run in Supreme Court of South Australia 2 7/25/10 Caught and convicted 13th INTERPOL Forensic Science Symposium, Lyon, France, October 16-­‐19 2001 FORENSIC EXAMINATION OF SOIL EVIDENCE In a murder case in California, vic/m’s body was dropped at a oil well apron where gravel which was transported from 300 mile south was used. Soil material found in the suspect’s car was compared with those around the oil well apron. The ques/oned sample from the car contained rock fragments which were the same with the imported gravel (1, 4). Blue thread gave key informa/on in a rape case in Upper Michigan. Three flower pots had been /pped over and spilled on the floor in the struggle. Podng soil on the suspect’s shoe was compared with one of those flower pot spillings. Small clipping of blue thread existed both in that flower pot sample and on the shoe of the suspect (1). Alterna/ve approach •  Profile the bio/c component of the soil •  Human ID = discrete en/ty •  Soil = small defined domain w/in a larger con/nuum –  Spa/o-­‐temporal dynamics need to be considered •  What is needed to use soil as evidence and/
or intelligence data? 3 7/25/10 COMMUNITY STRUCTURE •  Driven by soil type and environmental factors (Girvan, 2003) •  Microfauna present are indica/ve of the soil •  Sampled sites will disclose variability hDp://cropsoil.psu.edu/extension/livingmulch/images/9_soil_9lth.jpg FORENSIC SOIL ANALYSIS •  Soil characteriza/on –  Physical traits –  Chemical /elemental components •  Early 2000s concept arose: Horswell et. al –  DNA profiling of soil bacteria Terminal Restric/on Fragment Polymorphisms -­‐Experimentally Derived Lengths
G C
G C
G C
G C
G C
G C
G C
G C
4 7/25/10 PREVIOUS STUDIES Heath et. al (2006) Meyers et. al (2008) •  T-­‐RFLP method •  16S rRNA eubacterial gene •  T-­‐RFLP method •  16S rRNA eubacterial gene Conclusions •  Ecosystem-­‐discrimina/on Assump/ons –  Indicator TRFs •  Soil diversity exists •  Sample Homogeneity •  Temporal variability •  Within ecosystem clustering •  Temporal variability Moreno et. al (2005) • Sampled 3 soil types •  Wet/Dry season • Bacterial DNA profiling • LH-­‐PCR & CE • Mul/variate analyses 16S rRNA HYPOTHESIS H0: Soil bio/c communi/es do not vary among soil type 5 7/25/10 Moreno and others Prior studies – Queried only eubacteria •  Increase number of taxa assayed •  Maintain discrimina/on & increase resolu/on J Forensic Sci, November 2006, Vol. 51, No. 6
doi:10.1111/j.1556-4029.2006.00264.x
Available online at: www.blackwell-synergy.com
Lilliana I. Moreno,1,2 M.A., M.Fs.; DeEtta K. Mills,1,2 Ph.D.; James Entry,3 Ph.D.; Robert T. Sautter,2,w
M.S.; and Kalai Mathee,1,2 M.S., Ph.D.
Microbial Metagenome Profiling Using Amplicon
Length Heterogeneity-Polymerase Chain
Reaction Proves More Effective Than Elemental
Analysis in Discriminating Soil Specimens!
ABSTRACT: The combination of soil’s ubiquity and its intrinsic abiotic and biotic information can contribute greatly to the forensic field.
Although there are physical and chemical characterization methods of soil comparison for forensic purposes, these require a level of expertise not
always encountered in crime laboratories. We hypothesized that soil microbial community profiling could be used to discriminate between soil
types by providing biological fingerprints that confer uniqueness. Three of the six Miami-Dade soil types were randomly selected and sampled. We
compared the microbial metagenome profiles generated using amplicon length heterogeneity-polymerase chain reaction analysis of the 16S rRNA
genes with inductively coupled plasma optical emission spectroscopy analysis of 13 elements (Al, B, Ca, Cu, Fe, K, Mg, Mn, Na, P, S, Si, and Zn)
that are commonly encountered in soils. Bray–Curtis similarity index and analysis of similarity were performed on all data to establish differences
within sites, among sites, and across two seasons. These data matrices were used to group samples that shared similar community patterns using
nonmetric multidimensional scaling analysis. We concluded that while chemical characterization could provide some differentiation between
soils, microbial metagenome profiling was better able to discriminate between the soil types and had a high degree of reproducibility, therefore
proving to be a potential tool for forensic soil comparisons.
KEYWORDS: forensic science, soil forensics, microbial forensics, microbial profiling, amplicon length heterogeneity (ALH), soil metagenome,
inductively coupled plasma optical emission spectroscopy, elemental analysis
Soil is an ubiquitous material that can provide valuable clues in
forensic investigations due to the vast array of information contained therein. Crime scenes in which soil evidence was found
have established soils’ geological ability to provide critical findings and ultimately aid in the conviction or exoneration of individuals (1–4). However, soil analysis has been limited for the most
part to physical analyses of the material (5–7) that are conducted
by experts in the geological field. In addition, the type of analyses
performed usually requires large sample sizes that are seldom encountered in crime scenes (8). These analyses include, but are not
limited to, color, particle size distribution, microscopic comparisons, rock and mineral identification as well as chemical methodologies to identify fragments of glass and other trace materials
that might be found in the soil (9). We have limited this investigation to compare two methods, elemental and microbial analyses, that will require little or no extra expertise from the analyst
side. These two methods exploit an aspect of soil that, although
1
International Forensic Research Institute, Florida International University,
11200 SW 8th Street, Miami, FL 33199.
2
Department of Biological Sciences, Florida International University,
11200 SW 8th Street, Miami, FL 33199.
3
USDA Agricultural Research Service, Northwest Irrigation and Soils Research Laboratory, 3793 N 3600 E, Kimberly, ID 83341.
!Presented at Arthur M. Sackler Colloquium, National Academy of Sciences, Washington, DC; American Chemical Society, FAME, Orlando, FL;
Graduate Student Association Poster Competition, Florida International
University, Miami, FL; 2006 Annual Meeting of the American Academy
of Forensic Sciences, Seattle, WA.
w
Present address: USDA Subtropical Horticulture Research Station,
13601 Old Cutler Road, Miami, FL 33158.
Received 25 Mar. 2006; and in revised form 24 June 2006; accepted 1 July
2006; published & & &.
vastly used in the ecological field, has not been embraced in the
forensics field. Although it is very difficult and highly unlikely to
individualize soils because they are constantly in a state of flux, soil
comparisons for forensic purposes, given a known and an evidence
sample, can provide very useful, exclusionary information.
Elemental analyses of soils have been performed using sensitive techniques such as atomic absorption spectroscopy (AAS) and
inductively coupled plasma mass spectrometry (ICP-MS) to identify the elements in environmental samples (10–13). In this study,
inductively coupled plasma optical emission spectrometry (ICPOES) was used to measure elemental concentration of the sampled
soils. This technique, for example, has been used to identify heavy
metal contamination by providing information regarding the
forms of metals present in soil and has aided in establishing preventive and remediation techniques (14–17). These analytical
techniques have been frequently used with soils to obtain information that would otherwise be impossible to detect.
Except for the work of Horswell et al. (18), the biological aspect of soils for forensic purposes has been ignored. In that study,
they concluded that a given soil could be traced back to its origin
based on the microbial community profile obtained by terminal
restriction fragment length polymorphism (T-RFLP) analysis of
the 16S rRNA genes. This molecular profiling method has been
extensively used by ecologists to establish differences between
microbial communities (19–22). While it provides good discrimination, T-RFLP analysis requires a large amount of time, and the
output can be compromised by incomplete enzymatic DNA digestion (23). With the growing interest in microbial forensics and
the development of new techniques that have proven useful in the
forensic field, soil biotic characterization has been gaining import
and the wealth of information contained therein is being decoded.
SOIL SAMPLING SITES (2005) ULU-­‐2 1
Copyright r 2006 by American Academy of Forensic Sciences
FIU
ULU-­‐1 Moreno et. al 6 7/25/10 LH-­‐PCR All
16S
rRNA
genes
27F
355 R
Relative Intensity
V1-V2
All PCR!
products!
D.Mills Length (bp)
LENGTH HETEROGENEITY (LH-­‐PCR) ADVANTAGES •  Natural vs ar/ficially generated fragments •  No post-­‐PCR manipula/on •  Rapid, robust, reproducible method •  Could be used in most crime laboratories DISADVANTAGES •  Do not know what microorganisms are present •  Clone libraries and sequencing needed for defini/ve iden/fica/on to species level –  So does TRFLP and others 7 7/25/10 FOOD WEB APPROACH •  Community structure Plants Nematodes Soil Fungi Bacteria •  Mul/ple trophic levels •  Unique profiles 8 7/25/10 SOIL SAMPLE COLLECTION Soil Type (PBP or ULU) SP1 (A,B,C) SP2 (D,E,F) SP3 (G,H,I) DNA Extrac/on LENGTH HETEROGENEITY (LH-­‐PCR) •  Four “universal” primers sets •  Tested each primer set individually •  Op/mized two duplex PCR reac/ons – Bacterial/Fungal – Nematode/Plant PRIMER SELECTION PRIMER NAME TAXA (UNIVERSAL MARKER) TARGET REGION 27f EUBACTERIA 16S rRNA gene 355r EUBACTERIA 16S rRNA gene Ribosomal ITS1 NEM_ITS1f NEMATODE NEM_ITS1r NEMATODE Ribosomal ITS1 trnLf PLANT chloroplast gene trnLr PLANT chloroplast gene FUN_ITS1 FUNGI Ribosomal ITS 1 FUN_ITS2r FUNGI Ribosomal ITS 2 9 7/25/10 LH-­‐PCR PARAMETERS •  Single vs. Duplex •  Bacterial/Fungal –  28 cycles •  Nematode/Plant –  Step up 35 cycles DNA SEPARATION WITH ABI PRISM 310 GENETIC ANALYZER •  The DNA is separated in a single capillary through electrophoresis •  Electropherograms were generated and analyzed with GeneMapperTM research sotware, version 3.7 SINGLE VS. DUPLEX PROFILE Bacteria V1_V2 Single Reac/on Bacteria V1_V2 Duplex Reac/on 10 7/25/10 ANALYSES METHODS •  Mul/variate Data Transforma/on -­‐-­‐ Square root transformed –  Bray Cur/s Similarity – Bio/c data does not have normal distribu/on; no skew if amplicon is missing •  ANOSIM tests the null hypothesis Global R=0, no differences exist Global R=1, no similarity exists ANOSIM RESULTS Individual taxon Global R Fungal ITS1
0.208 Nematode ITS1 0.370 Bacteria V1_V2 0.424 Plant trnL 0.554 ANOSIM RESULTS Combina/on of markers Global R Combina/on of markers Global R Bacteria/Plant 0.251 Fungal/Plant 0.369 Bacteria/Fungal 0.424 Bacteria/Nematode/Plant 0.619 Bacteria/Nematode 0.677 Bacteria/Fungal/Nematode 0.431 Nematode/Plant 0.769 Bacteria/Fungal/Plant
Nematode/Fungal
0.350 0.438 Bacteria/Nematode/Plant/Fungal 0.663 11 7/25/10 NONMETRIC-­‐MULTIDIMENSIONAL SCALING (MDS) Bacteria V1_V2 similarity MDS RESULTS Fungal ITS similarity Plant trnL similarity Nematode ITS similarity 12 7/25/10 FOUR TAXA COMBINED SIMPER METHOD •  Similarity Percentages – iden/fies variables (amplicons) driving dissimilari/es between subplots and/or sites Overall 80% Dissimilarity SIMPER RESULTS AMPLICON (BP) AVERAGE AVERAGE AMPLICON ASSOCIATED AMPLICON AMPLICON PERCENT TAXA ABUNDANCE ABUNDANCE DISSIMILARITY
(ULU) (PBP) (cumula/ve%) 137 0.07 0.49 2.91 129 0.00 0.39 5.56 P F 139 0.39 0.00 8.21 N 339 0.00 0.34 10.68 B 153 0.33 0.00 13.07 P 114 0.00 0.33 15.32 N 124 0.26 0.00 17.16 N 150 0.05 0.27 18.90 P/N 341 0.27 0.00 20.63 B 120 0.17 0.26 22.34 N 177 0.25 0.00 24.05 N/F/P 340 0.21 0.08 25.75 B 13 7/25/10 BACTERIAL VS METAGENOMIC DNA PROFILING OF SOIL COMMUNITIES •  Bacteria do discriminate •  Metagenomic DNA discriminates as well – All 4 taxa contribute to differences – “Uniqueness” APPLICATIONS •  Forensic Laboratories –  Trace evidence –  Bioterrorism •  Intelligence •  (geographic origin) •  Microbial Ecology Field –  Bioremedia/on –  Seasonal/Temporal Varia/on DTD 5
ARTICLE IN PRESS
Journal of Microbiological Methods xx (2005) xxx – xxx
www.elsevier.com/locate/jmicmeth
An ecoinformatics tool for microbial community studies:
Supervised classification of Amplicon Length
Heterogeneity (ALH) profiles of 16S rRNA
Chengyong Yang a, DeEtta Mills b, Kalai Mathee b, Yong Wang a, Krish Jayachandran c,
Masoumeh Sikaroodi d, Patrick Gillevet d, Jim Entry e, Giri Narasimhan a,*
a
Bioinformatics Research Group (BioRG), School of Computer Science, Florida International University, Miami, Florida, 33199, USA
b
Department of Biological Sciences, Florida International University, Miami, Florida, USA
c
Department of Environmental Sciences, Florida International University, Miami, Florida, USA
d
Microbial and Environmental Biocomplexity, Department of Environmental Sciences and Policy, George Mason University,
Manassas, Virginia, USA
e
USDA Agricultural Research Service, Northwest Irrigation and Soils Research Laboratory, Kimberly, Idaho, USA
Received 18 January 2005; received in revised form 22 April 2005; accepted 24 June 2005
Abstract
Support vector machines (SVM) and K-nearest neighbors (KNN) are two computational machine learning tools that
perform supervised classification. This paper presents a novel application of such supervised analytical tools for microbial
community profiling and to distinguish patterning among ecosystems. Amplicon length heterogeneity (ALH) profiles from
several hypervariable regions of 16S rRNA gene of eubacterial communities from Idaho agricultural soil samples and from
Chesapeake Bay marsh sediments were separately analyzed. The profiles from all available hypervariable regions were
concatenated to obtain a combined profile, which was then provided to the SVM and KNN classifiers. Each profile was
labeled with information about the location or time of its sampling. We hypothesized that after a learning phase using
feature vectors from labeled ALH profiles, both these classifiers would have the capacity to predict the labels of previously
unseen samples. The resulting classifiers were able to predict the labels of the Idaho soil samples with high accuracy. The
classifiers were less accurate for the classification of the Chesapeake Bay sediments suggesting greater similarity within the
Bay’s microbial community patterns in the sampled sites. The profiles obtained from the V1 + V2 region were more
informative than that obtained from any other single region. However, combining them with profiles from the V1 region
(with or without the profiles from the V3 region) resulted in the most accurate classification of the samples. The addition
* Corresponding author. Tel.: +1 305 348 3748; fax: +1 305 348 3549.
E-mail address: [email protected] (G. Narasimhan).
0167-7012/$ - see front matter D 2005 Elsevier B.V. All rights reserved.
doi:10.1016/j.mimet.2005.06.012
MIMET-02319; No of Pages 14
14 7/25/10 Support Vector Machines Training set vs unknowns using LH-­‐PCR (all) concatenated data Single 16S domains 15 7/25/10 Tested three domain combina/ons Chesapeake Bay sediment samples V1 +V2 domain only Conclusion •  Get to know your computer science colleagues!! •  Large databases need to be established with profile data and then unknowns can be classified as to where and when they were ‘collected’; no longer a ‘crap shoot’ 16 7/25/10 What is needed to use soil as forensic evidence? The 2nd Soil Forensics International Conference
http://www.soilforensicsinternational.org/sfi2010.php
Soil Forensics International
Home
SFI 2010
Post-event 2007
Posters &
presentations
Programme
Criminal topics
Environmental
topics
Keynote Speakers
SFI 2010
3rd International Workshop on Criminal
and Environmental Soil Forensics
The dirty evidence: soil and geoforensic
contributions to intelligence gathering and
environmental and public safety
Media Coverage
Referring sites
Organising
committee
Sponsors
The next international meeting on soil forensic
analysis and investigation will be held in Long Beach,
California, USA, from 31st October 2010 to 4th
November 2010.
Sponsored by: California Association of Criminalistics,
California Department of Justice, and Soil Forensics
International
Further information is available at:
www.acsmeetings.org/
Topics include:
1. The use of forensic geoscience in the intelligence
community
2. Legal issues facing forensic geoscience practitioners
1 of 3
7/25/10 11:47 AM
Same thing as all other forensic disciplines? •  Expert witnesses/exper/se in the field –  Balance of fundamental soil science with forensics interpreta/on •  Acceptability of methods –  But: no ‘soil standards’, no SWG-­‐SOIL, no training, no common SOPs, no standard sta/s/cs, no proficiency tes/ng •  Legal considera/ons –  Daubert, Frye standards met? Class par9cipa9on! 17 7/25/10 Ques9ons? “Soil mapping is possible only because men can examine a profile at one point and successfully predict its occurrence at another point where surface indica9ons are similar.” -­‐-­‐-­‐ Author unknown 18