\n\n

No Positive Food Sample? Whole Genome Sequencing and Traceback Still Tell Us Who Got Sick and Why

by | Nov 2, 2026 | Law | 0 comments

Here we go again

I have been doing this since the Jack in the Box E. coli outbreak in 1993. In the thirty-plus years since, I have heard one defense argument more than any other: “Nobody ever found the bacteria in our food.”It sounds like a good argument. It isn’t. It misunderstands how CDC, FDA and state health departments actually solve outbreaks, and it ignores the most powerful tool public health has ever had for linking sick people together: whole genome sequencing, or WGS.

Here is the short version. WGS tells us which sick people are part of the same outbreak. Epidemiology tells us what they ate. Traceback tells us where that food came from. When those line up, the government names the food, positive food sample or not. And when the traceback from ill people converges on a single farm, plant or supplier, the case gets even stronger.

WGS is DNA evidence for bacteria

Think of WGS as forensic DNA testing for germs. A public health lab takes the Salmonella, E. coli or Listeria cultured from a sick person’s stool or blood and reads essentially its entire genetic code. Isolates that are nearly identical share a recent common source. Period.

CDC’s PulseNet network uses WGS to build those DNA fingerprints and then looks for clusters of matching patterns that signal an outbreak. PulseNet retired the older PFGE method and made WGS its primary surveillance tool, in part because PFGE could lump unrelated strains together. FDA runs a companion network, GenomeTrakr, that sequences bacteria from food, farms and plants and posts them to a public database. A multiagency federal report describes the two systems as complementary tools for linking related pathogens during outbreaks.

Here is what matters in a courtroom: the government’s outbreak case definition is a WGS definition. In the 2022 Wendy’s E. coli outbreak, a case was a person whose isolate was closely related to the outbreak strain by cgMLST, with illness onset in a set window. If your bug matches, CDC has already decided you are part of the outbreak.

And WGS is honest. In the 2015 Blue Bell Listeria outbreak, it added three Texas patients and kicked out a fourth whose isolate did not match. A test that can say “no” is a test whose “yes” means something.

The food is usually gone

A negative food test usually means the contaminated lot is long gone, not that it was ever clean.

  • Time. It typically takes 3 to 4 weeks to determine whether a sick person is part of an outbreak. Lettuce, onions and restaurant ingredients turn over in days.
  • Shelf life. In the 2023 pico de gallo Salmonella outbreak, by the time FDA had enough to name the product, it was past its shelf life. Same with the sprouts in a 2025 E. coli O26 outbreak.
  • Sampling. Contamination is spotty and low-level. A few samples from a later lot or a freshly sanitized plant can easily miss it.
  • Foot-dragging. In the 2018 Yuma romaine outbreak, romaine was implicated in mid-April, but canal water was not sampled until June 4–8. When it finally was, it matched.

CDC admits it doesn’t solve every outbreak because some end before enough information is gathered. Food testing is one of three kinds of evidence it uses, not the price of admission.

How the government actually solves outbreaks

CDC uses three types of data to find an outbreak food: epidemiologic, traceback, and food and environmental testing. It does not need all three, and it names sources on epidemiology and traceback alone all the time. It works like this:

  1. WGS links the people. Genetically matching isolates tell investigators the patients share a source, even if they live in different states and never met.
  2. Epidemiology points to the food. Interviews show what those patients ate, and whether one food shows up far more often than it would by chance.
  3. Traceback points to the source. Investigators follow individual patients’ purchases back up the supply chain, looking for a common point in the distribution chain where the food could have been contaminated.

Now think about what convergence means. Patients in different states, who bought food at different stores and restaurants, all trace back to one grower or one processor. That is not coincidence. That is the answer.

The CIFOR guidelines describe investigations the same way: epidemiology, traceback and environmental assessment, and lab data, integrated. And CDC says it acts when there is clear and convincing information linking illness to a food. That is a higher bar than the more-likely-than-not standard my clients have to meet in a civil case.

The examples, and there are plenty

Every outbreak below was solved by WGS, epidemiology and usually traceback. In none of them did anyone find the outbreak strain in a food sample.

Outbreak Pathogen Illnesses What tied illnesses to the food Food and site testing
McDonald’s slivered onions (Taylor Farms), 2024 E. coli O157:H7 104 in 14 states; 34 hospitalized, 1 death (CIDRAP) 80 of 81 interviewed ate at McDonald’s; 63 ate an item with slivered onions; traceback to Taylor Farms Outbreak strain never found; one onion and one grower sample had a different, non-matching STEC
Grimmway Farms organic carrots, 2024 E. coli O121 48 in 19 states; 20 hospitalized, 1 death (FDA) 26 of 27 interviewed ate carrots; traceback found Grimmway as the common supplier Two environmental STEC positives did not match (summary)
Pico de gallo, 2023 Salmonella Paratyphi B 37 Epidemiology and traceback All samples negative; product past shelf life
Indiana cantaloupe, 2022 Salmonella Typhimurium 88 in 11 states; 32 hospitalized Epidemiology and traceback to southwest Indiana melons Salmonella found on farms and in the packinghouse; none matched
California peaches, 2020 Salmonella Enteritidis 101 in 17 states; 28 hospitalized Epidemiology and traceback to one large grower No sample matched
Sprouts, 2025 E. coli O26:H11 10 Interviews and traceback None available; past shelf life
Evergreen Fresh sprouts (Idaho), 2011 Salmonella Enteritidis 25 in 5 states 69% of early cases ate sprouts vs. a 4.4% background rate; traceback to one grower Everything tested negative

McDonald’s is the textbook case. FDA said flatly that the outbreak strain was not found in onion or environmental samples, yet epidemiology and traceback showed the recalled onions were the likely source. Taylor Farms recalled. McDonald’s switched suppliers. New cases stopped once the Quarter Pounder came off the menu. Grimmway played out the same way a month later.

Even when the ingredient stays uncertain, WGS still ties the victims to the place. In the 2022 Wendy’s outbreak, WGS identified 109 sick people in six states. More than 80% of those interviewed ate at Wendy’s, and 46 of 68 with detailed histories ate romaine on sandwiches. Traceback never converged on one grower, so CDC did not confirm the romaine. Nobody seriously argued the illnesses came from somewhere other than Wendy’s. And the outbreak strain was closely related to a 2021 spinach outbreak strain, pointing right back at leafy greens.

When the match finally shows up, it lands right where WGS said it would

Sometimes investigators do eventually find the outbreak strain in a field, a canal or a plant. When they do, it confirms what WGS and traceback already told us, sometimes years earlier.

The match was never what made the case. It just proved the science right, again.

What this means for people who got sick

If your isolate is a WGS match to an outbreak strain, and epidemiology and traceback point to a food, you have the same proof the government uses to order recalls. Demanding a positive leftover on top of that asks more of a sick child’s family than CDC and FDA ask of themselves.

Courts get this. In Patterson v. Kevon, LLC (Ga. 2018), the Georgia Supreme Court rejected the idea that food poisoning plaintiffs carry a heavier summary judgment burden than anyone else. In Goldman v. Chop’t Creative Salad Co. (N.Y. App. Div. 2024), negative supplier tests and no other sick diners were circumstantial and not dispositive, and the case moved forward.

Those cases did not even involve an outbreak. WGS adds what older cases lacked: objective, molecular proof that my client’s bacteria came from the same source as dozens of others. When those others all trace back to one supplier, the defense is left arguing that genetically identical infections across multiple states share some cause other than the one food everyone ate.

If you get sick in an outbreak:

  • Get a stool test. No isolate, nothing to sequence.
  • Talk to the health department. Your food history and receipts drive the epidemiology and traceback.
  • Keep everything. Leftovers, packaging, receipts, loyalty card records. A positive sample isn’t required, but it never hurts.

The genome doesn’t lie. Neither does a traceback.

Sources

Submit a Comment

Your email address will not be published. Required fields are marked *