This video was producedby the Food and Drug Administration.

.

.

.

 Food.

.

.

so enjoyable, so delicious, such asatisfying and essential part of our lives.

.

.

andyet, do most of us ever wonder whereit comes from? Where it's grown? How it's handled andpackaged and shipped? Nope, most of us don'tthink about any of that - not until something we'veeaten makes us sick.

 And then, how do we figureout which particular food was the culprit? Think about a salad bar,where you might easily eat 50 different ingredients.

 Each different kind offood can come from a different part of theworld, and each one has different pathogensassociated with it.

.

.

and that means that every timeyou enjoy a salad, you're really exposing yourselfto a large number of global pathogens, any ofwhich have the potential to make you sick.

 According to the Centersfor Disease Control, foodborne diseaseoutbreaks are responsible for about 48 millionillnesses, 128,000 hospitalizations, and3,000 deaths in the United States every year.

 As the world becomesmore interconnected, maintaining a safe foodsupply has become an increasing challenge,and the Food and Drug Administration haspioneered an innovative response tothis challenge.

 Collaborating with federaland state public health labs, FDA has built apathogen identification network calledGenomeTrakr.

 It uses whole genomesequencing to look at the DNA fingerprint leftbehind by disease-causing bacteria, allowing FDA toidentify the source of an outbreak with more detailand clarity than ever before.

 Whole genome sequencing ismuch like a radar gun to a police officer or statetrooper who is looking for speeders.

 Whole genome sequencingwill now be able to surveil the foodsupply for pathogens.

 Much like a radar thatruns from the surface of the ocean to the edge ofspace, we'll be able to detect outbreaks when theyhappen early and we'll be able to stop them in theirtracks before they become a public health crisis.

 The Genometrakr networkhas three different elements to it.

.

.

.

 The first is just tobuild a database of known pathogens from known foodsfrom known geographic localities.

 We know from the examplesthat we've tested that you can tell salmonella,whether it's from California, or from China,and that information can help an investigation indetermining where the contaminants enteredinto the food supply.

 So at FDA, of course,we have a very large surveillance program andwe, of course, test many foods that come in atports of entry and we test many domestic foods beforethey go into the food supply.

 Sometimes we findpathogens like salmonella or e-coli or listeria.

 And when we do, we nowsequence them to see if they're related to a foodborne outbreak event, and then more so to see if wecan track it back to the very source fromwhere it came.

 Pathogens evolve veryquickly and have thousands of genetic variations.

 After spending time ina particular geographic location, a pathogen likeSalmonella begins to acquire unique geneticsignatures, like fingerprints, thatidentify it as coming from that location.

 One good example of wherewhole genome sequencing has shown its reallyunique power was in 2012, when we had an outbreakrelated to spicy tuna associated with sushi.

 So we knew people weregetting sick, and we knew that some of them saidthat they'd eaten at a sushi place, but we didn'tknow which ingredient was responsible.

 In fact, initially theythought it was in the hot sauce.

 And so what we did was, wesequenced a dozen isolates from our freezer from foodfrom around the world that were related to thisparticular bacterium.

.

.

.

 When FDA compared thebacteria from the patient back to the samples inGenomeTrakr's database, the culprit turned outto be a pathogen called Salmonella Bareilly.

.

.

.

.

 And unfortunately,Salmonella Bareilly can be found all over the easternhemisphere of the world in many different places, andour current forensic tools were not able todistinguish the source.

 When we applied wholegenome sequencing to those isolates of salmonellathat we collected from all over the world, we wereable to link it directly back to reference isolatesthat came within six miles of the actual tunafacility in Southeast Asia where the firm was thatwas contaminating the food supply.

 This was a major findingbecause it showed that pairing genomicinformation with geographic informationcould narrow the search for the source of acontaminated ingredient, even when the source waslocated halfway around the world! We never knew thatsequencing isolates from a port of entry that werecollected five years before that event, wouldtell us the potential source of the outbreak.

 And whole genomesequencing is also telling us things we never thoughtwe could learn about bacterial pathogens.

.

.

.

.

 .

.

.

Things like certainsalmonella are now resistant to heavy metalsand they live in tuna and other top predators in thesea where heavy metals accumulate.

 These are adaptations orchanges in the germ that we never knew existed.

 And by learning how thegerm itself is evolving and changing, we'll beable to come up with better preventive controlsfor the food supply and for food manufacturers tostop the contamination in their plants, in theirfacilities before it spreads.

 In 2014, the GenomeTrakrnetwork played a role in a regulatory action by theFDA and traced the source of a multi-stateListeriosis outbreak right to the firm that made thecontaminated product - a Queso-style cheese.

 Their pasteurizationprocess was not working correctly, so the cheesewasn't being sterilized, if you will, to keepthe pathogens out.

 And once the listeriaestablished itself in the facility, it wasin everything.

 And so almost every batchof cheese was contaminated with listeria, so anyonethat had a particular susceptibility to thisorganism was getting sick.

 FDA was able to sequencethe isolates of listeria from cheese collected fromthe firm and compare them with isolates that CDC hadcollected from the sick patients.

 Typically what happensis people get sick, they present themselves toa physician and the physician says oh, youhave disease X, Y or Z, in this case listeriosis.

 It's fairly rare, sophysicians know to upload the information.

 In this case, we wouldsequence it and then you start comparingthe information.

 What was interesting aboutthis case is, in the past, this would've occurredover a series of months.

 Now, because of thedigital sharing of information, we were ableto move much faster with this type of technology.

 And so we can now say oh,well instantly we have a match, much like acriminal case where you have DNA evidence, and youcan link a criminal to a particular scene, we wereable to link a food to a group of people thatwere getting sick.

 And this was the firsttime this was done.

 Whole genome sequencingstrengthened FDA's evidence linking theoutbreak strain to these cheese products.

 FDA was able to shutthe facility down.

 And then a differentoutbreak occurred.

.

.

.

.

 We noticed that there wereillnesses associated with peanut butter or nutbutter, and they were in different states anddifferent parts of the country and it was notclear that they were related.

 At the same time FDA fieldinspectors had done a survey of a plant thatmade nut butter and found salmonella inthe facility.

 When we sequenced thosesalmonella and put them in the database, they werealmost perfect matches across 4.

6 million datapoints to the illnesses in different partsof the country.

 So we had actually used itin this case to detect the source of an outbreakjust as the outbreak was starting.

 We were able to stop thefood product from going further into theconsumer's food supply, and we were able to stopthe outbreak before it advanced to many dozensor hundreds of illnesses.

 The second element ofGenomeTrakr is collecting Global samples.

.

.

 Getting federal, state andinternational partners collaborating to build thedatabase, because bacteria are globally distributedand food is globally traded.

 So any food you might buyfrom your local grocery store or eat at a localrestaurant may have come from anywhere around theworld, and so we need global collectionsto also be included.

 Surprisingly, it doesn'treally take that much time to collect all thisgenomic data.

.

.

.

.

 It takes a little lessthan a week and the process has manycomponents.

 The first is that actuallyit's isolating the bacteria from the actualfood, so for example, the way we soak a tomato andget the bacteria out of it, is different than howwe treat avocado or peanut butter.

 There's a whole group ofthe FDA that specializes in how best to getpathogens out of different kinds of food.

 Once you have thepathogen, then it's relatively straightforward to produce DNA and then do the chemicalreactions to build a genomic library tosequence that, and then there's a process ofthe data analysis.

 We move the data from thesequencer into a high performance computer andthen we do data analysis and then give thoseresults to a subject matter expert who canassess the results and provide an interpretation.

 And the third element issharing all this data, which is housed at TheNational Institutes of Health.

.

.

.

.

 Data is publiclyavailable, anybody can see the data, screen the dataand can compare their particular bacterialpathogen to the database to see, then characterizewhat pathogen they have discovered.

 Not only are we making thedata and metadata public in real time, we are alsomaking the data analysis public.

 Our partners at NIH arebuilding an analysis pipeline off the databaseso the genome data comes in, the new genomes arematched against the existing database and atree is built revealing the evolutionaryrelationships of all the bacteria in the database.

 As you can imagine,the tree is huge right now.

.

.

.

.

and so when a newisolate comes in, it's like a new twig beingadded to the big evolutionary treeof pathogens.

 The applications of thisare probably something that we can't evenforesee right now.

 But the actual genes andthe actual variation that is arising in thesepathogens might be used for many other usefultools, for example, new vaccines, or newunderstanding of pathogens, new ways tocontrol bacteria to get it out of, say, a farm or amanufacturer's facility.

 By making the datapublicly available, we allow a larger group ofpeople to see it and explore it and then comeup with new ways to use it.

 And the FDA really doeswant a larger group of people involved.

 They're very interestedin finding new state, national, andinternational partners.

.

.

.

.

 We can help in many ways.

 We have a large set ofresources as well as the experiences of setting upa network, and so we are happy to help you howeveryou would like us to.

 That may be education,that might be sequencing isolates or that might bejust integrating you in our data analysis process.

 We're happy to work withyou in any way that we can help.

 GenomeTrakr is just in itsinfancy, but already it has had a profoundimpact on food safety.

 It began with a databaseof information on pathogens like salmonella,listeria and e coli, but in the years ahead, itwill include data on any bacterial or viralpathogen, not only foodborne pathogens butinfectious agents like tuberculosis, SARS,hospital-acquired infections and eventhose affecting animals.

 By bringing whole genomesequencing into the laboratory, FDA, alongwith its growing numbers of domestic and globalpartners, is truly pushing back the frontiers ofoutbreak response.

.

.

.