How CSI Wildlife Actually Works for Ivory Tracking
The CSI Wildlife project from Cold Spring Harbor Laboratory is a bioinformatics teaching tool that lets students match confiscated ivory to poached elephant populations using DNA barcoding. The platform uses mitochondrial DNA sequences, specifically a 426-base-pair region of the cytochrome b gene, and compares them against a reference database of known elephant populations across Africa. You load a sequence, run a BLAST search against the provided database, and determine the geographic origin based on the closest genetic matches. It is straightforward in concept but has enough quirks to frustrate people who have walked through it more than once. Most people searching for an answer key are working through the lab modules on the Learn.Genetics site. The exercises are structured in two main parts: one that focuses on identifying the species of seized ivory (elephant versus other animals) and a second that pins down the poaching location within Africa. The answer key you end up needing usually covers four or five specific labs, each with sequence data files and multiple choice or short answer questions. I spent a semester helping undergraduate students navigate this material, and the most common sticking point is not the biology itself but the mechanics of using the BLAST tool inside the simulation. Students will paste a sequence incorrectly, include whitespace characters from a PDF copy, or accidentally trim a base from the wrong end. I once had a student who spent twenty minutes confused about why every result came back as African bush elephant when she was supposed to be analyzing a sequence that should have matched a forest elephant. The issue was a single extra character at the 5' end that she had picked up when copying the sequence from a screenshot rather than the text file. No amount of re-reading the instructions fixed that. She had to download the raw FASTA file directly.
The labs themselves walk you through loading sequences into BLAST+, running a nucleotide BLAST against the elephant cytochrome b reference set, and interpreting the percent identity and E-value results to assign a geographic origin. The reference database contains samples from fourteen countries, and the sequences cluster in ways that reflect real population structure. Southeast Asian elephant sequences show up in some labs too, which adds a species discrimination step before the geographic matching even begins.
Running Through the Lab Modules
Module one asks you to determine whether unknown ivory comes from elephant or another species. You get three sample sequences and you run each against NCBI's GenBank. Elephant sequences return nearly perfect matches with E-values near zero and percent identities above 98 percent. The trick is that some marine mammal or rodent sequences can return deceptively high similarities in a short fragment, so you need to check both the taxonomic classification and the alignment length. A 95 percent match over only 50 bases means something very different than a 95 percent match over the full 426 base pairs. Module two is where the geographic assignment happens. You are given sequences from seized ivory and you compare them against the curated reference database that the platform provides. The database is built from historically seized ivory and blood samples collected across the continent. When you run BLAST, the top hits will usually point to a specific country or region. The answer key questions then ask you to interpret what those results mean in conservation terms: which populations are under the most pressure, which corridors show the highest movement of ivory, and so on. One thing the answer key does not always make clear is how to handle ambiguous results. Sometimes a sample will have nearly identical matches to two different regions, or the top hit will be a country with very few reference samples. In those cases the platform expects you to note the uncertainty and discuss the limitations of the reference database rather than force a single confident answer. I learned this the hard way when a student told me she was stuck because two countries tied for first place. She thought she had made an error. She had not. That is what the real data looks like.
Get the Full Details
Common Pitfalls and Workarounds
The platform runs best in Chrome or Firefox. Safari has caused intermittent issues with the embedded BLAST interface, and some students report that their BLAST jobs time out if they leave the tab idle for more than ten minutes. The simulation does not always save your progress if the browser crashes, which is annoying when you are halfway through a multi-step analysis. Another issue is the formatting of the answer key questions themselves. Some versions of the lab ask you to submit specific sequence IDs or accession numbers, and if you copy those from the BLAST results page you can pick up invisible formatting characters. I found that manually retyping the accession number rather than copying it eliminated a whole class of submission errors for my students. The system accepts both formats, but the automated grader does not always strip hidden characters the way a human would. The sequence files you download are in FASTA format. Make sure you are using the version labeled with the full 426 bp fragment. There are shorter variants in the database and if you accidentally upload one of those, your BLAST results will look different and your answers will be wrong. The difference is subtle and easy to miss at first.
Understanding the Science Behind the Answers
Before you look at any answer key, it helps to understand what the cytochrome b gene actually tells you. It is a mitochondrial marker, which means it is maternally inherited and does not recombine. That makes it useful for tracking population structure because it accumulates mutations at a steady rate over generations. The geographic clustering you see in the reference database reflects historical migration patterns and current population fragmentation, not individual animal movement. When a seized ivory sample matches a reference from Central African Republic, that does not mean the elephant was killed there. It means the genetic profile is consistent with populations in that region, and given the trade patterns documented in the literature, that is the most likely origin. The reference database has known gaps. Some countries have dense sampling and others have very little. The western African forest elephant population is particularly understudied genetically, so ivory from that region sometimes returns ambiguous results or matches to the nearest sampled population rather than the true origin. The answer key questions acknowledge this in later modules, but if you are working through the first pass you might not realize why some of your assignments feel unsatisfyingly imprecise. A counter-intuitive point that trips up a lot of students is the relationship between genetic distance and geographic distance. These elephants are not organized into neat national bins. The gene flow across borders means a sample from one country can be genetically closer to a population fifty kilometers away in a neighboring country. The answer key sometimes frames results in terms of countries because that is how enforcement agencies think about it, but the underlying biology does not respect those boundaries. A good answer acknowledges both the practical and the biological reality.
Where to Find the Materials
The labs are hosted on the Learn.Genetics website at the University of Utah, and they are freely available without registration. The sequence data files are provided as downloadable attachments within each module. There is no separate official answer key document published by the developers, which is why so many students end up searching for one. The answers are embedded in the instruction text and discussion prompts of each lab module, and the instructors who have used this curriculum typically work from their own answer sheets that they develop from the data. If you are a student working through the materials independently, the most reliable approach is to run the BLAST searches yourself and compare your results against the expected outcomes described in the lab instructions. The questions are designed so that the data leads you to the answer. If you are getting results that do not match the discussion points, the problem is almost always in the input step rather than in your interpretation of the output.
What This Tool Actually Misses
It is worth being honest about the limitations. CSI Wildlife is an educational simulation, not a forensic-grade tool used in actual poaching investigations. Real wildlife forensic labs use a combination of mitochondrial DNA control region sequencing, nuclear microsatellite markers, and increasingly whole-genome approaches. The cytochrome b marker used here has limited resolution compared to those methods. It can tell you the broad region of origin but cannot pinpoint a specific herd or individual. For educational purposes this is sufficient, but anyone planning to apply these techniques in real conservation work needs to understand that the resolution is coarse. The database itself is a snapshot in time. New poaching outbreaks generate new genetic signatures that may not yet be represented. Ongoing sampling efforts are filling in gaps, but the coverage remains uneven. If you are using this for anything beyond the classroom, you should be citing the publication that describes the reference dataset and noting its limitations explicitly. The original research behind this tool was published in PLOS Biology by Goodman et al., and it remains the best primary source for understanding what the database can and cannot do. If you are writing a report or preparing materials, referencing that paper will give you more accurate context than any summary answer key ever could.
