How much of Thermo Fisher's antibody data has been manipulated? | Hacker News Reader

How much of Thermo Fisher's antibody data has been manipulated?

reeserichardson.blog

447 pointsmhrmsn91 comments

My first, second, and third instinct here is to say this is pretty obvious and sloppy fraud. But it did remind me of the famous case discovered by David Kriesel where Xerox scanners changed documents in surprising ways. The caption on the YouTube video linked here is entertainingly accurate.

https://www.dkriesel.com/en/blog/2013/0802_xerox-workcentres...

"On the scale of things too horrible to contemplate, "document-altering scanner" is right up there with "flesh-eating bacteria". Since 2006, Xerox scancopiers literally are making stuff up. They, for example, replace digits with others in scans. The replacement digits are layouted perfectly into the page, so the errors are hard to see. Sounds unbelievably insidious, but it's true. Drug prescriptions, construction plans, just anything can be affected. "

And now practically every phone camera "enhances" the image via AI that might invent details. Most famously Samsung adding details to photos of the moon.

I'm scared that the "AI" marketing will make it much harder convincing non-technical coworkers and execs that "garbage in, garbage out" is a real concept, that not all "data" is good, and that our systems need to keep track of which kind is going where.

"All data is useful, the more the better! Just put it all into the AI and it'll sort it out."

Is this really pervasive? E.g. To my knowledge the "AI" enhancement that iPhones do automatically is limited to the usual sorts of post-processing for contrast, color, etc. There is an AI editing mode that leans more into generative fill capability that would be analogous to the Samsung incident but I don't think it's happening automatically to every photo you take.

I still remember Samsung faking images when using digital zoom...

https://www.theverge.com/2023/3/13/23637401/samsung-fake-moo...

Yes, I believe that is what the person I replied to was referring to specifically.

We faced this problem recently when doing an old-fashioned experiment to test a hypothesis. Can you trust the images from your phone to check if phenomenon X has occurred or not? What if the AI model processing the image just hallucinated what you wanted to see?

It is Descarte's evil demon incarnate. Yet another incentive to preserve my old mirrorless camera.

Nice! Just because you’re the top comment, I feel it is helpful to quick readers to point out that it has 0 to do with what happened here. :)

> They, for example, replace digits with others in scans. The replacement digits are layouted perfectly into the page, so the errors are hard to see.

I dealt with this where our fax number had a 6 in it and it would sometimes get changed into an 8, which happened to be a valid fax number for another company, ugh. And this was confidential info too…

Always a funny phone call when they insist they sent it to the number on the cover page we sent and then they send us a copy and xerox made it wrong.

> They, for example, replace digits with others in scans. The replacement digits are layouted perfectly into the page, so the errors are hard to see.

Is this with JBIG2? I remember reading about JBIG2 also used in the FORCEDENTRY zero click exploit that which was (?) used in the Pegasus spyware. Unrelated tidbit, I guess.

The Xerox problem was with image compression. The compressor would look at an incoming block, decide it was similar enough to one already noted and reuse the existing token. With small fonts and similar characters one difference was within the tolerance and after seeing something like 888 it could then decide 8B8 was the same thing. As it could only happen when things lined up perfectly the alterations would also be lined up perfectly.

Yes, this algorithm is called JBIG2.

Yes.

The guy who uncovered this, Sholto David, is basically just awesome?

Watch him cycle from Wales -> China in 90 days: https://www.youtube.com/watch?v=MdgHZPfivVA

This isn't his first fraud rodeo either. For his discovery of serious fraud by the Dana-Farber Cancer Institute in 2024, he received $2.6 million.

Be more like Sholto, exercise your free will!

I didn't realize whistleblowers could recover part of fraud settlements:

> The civil settlement includes the resolution of claims brought under the qui tam or whistleblower provisions of the False Claims Act by Sholto David. Under those provisions, a private party can file an action on behalf of the United States and receive a portion of any recovery. David will receive $2,625,000 under today’s settlement.

https://www.justice.gov/opa/pr/dana-farber-cancer-institute-...

https://www.nbcnews.com/news/investigations/marine-sarah-fei...

$40m for this person in the US

worth it for a lifetime of looking over your shoulder? maybe...

If you believe the conspiracy theories, that Boeing whistleblower “took his own life” before testimony and any potential payout.

Short sellers have become the biggest SEC whistleblowers if they discover suspected fraud - it’s a more lucrative and less risky than shorting:

https://www.institutionalinvestor.com/article/2d7d14lmtfwc1e...

Gather round kids and grab your muskets. The Qui Tam Clan ain't nuthin to fuk wit.

Are you sure?

Did he donate the money away? In the video you sent, he seems quite anxious about his nightly budget.

EDIT: ah it seems his trip was before that

This is systematic fraud, and anyone trying those antibodies with falsified data will waste money and time. A lot of papers have been retracted for similar issues. Thermo Fisher is a major worldwide supplier of antibodies, so this has quite a big practical impact.

1. Obviously unethical and fraudulent behavior.

2. It should be determined whether the fraud was just the display image (imagine a sales manager making a bad call when images are not available) or involved the underlying research (more systemic and worrying).

3. It would be interesting to examine occurrence of faked images along with apparent unreliability/irreproducibility of research that has used those products.

There aren't any display images in science.

Manipulating images for presentation is an automated process unless you're ripping someone off. The changes would be uniform across whole sets.

The problem with trying to pass off a fake image is that you need to be more knowledgeable in each dimension of the effort than the recipients are in just one. If anyone remembers the folks identifying East German video from background hum it's kind of like that.

"a sales manager making a bad call when images are not available"

It seems nearly impossible to imagine that to be the case. I'd have to disregard the kinds of manipulation entirely. What sales manager would create a whole western block sequence by copying, rotating, and flipping a single element?

Someone else here mentioned proofers, people that prepare marketing materials, as a potential source. I am in no way defending Thermo here. I just meant that the extend of the fraud needs to be determined, from some non-scientist making a decision for short term profits, frustrated that no one saved the picture, or because the pictures showed how ugly the western blots actually are, versus wholesale fabrication of the research from the bottom up.

> people that prepare marketing materials, as a potential source

Scientific advertising and marketing is a small, specialized field, done by people with fairly solid technical backgrounds (we produce a whole lot of advanced STEM degrees, there's plenty of folks available with this sort of background).

So I just want to be crystal, crystal clear here: there's no way in hell anyone involved in this pipeline should have any confusion as to whether "improving" gel photographs by painting out details and/or copying and pasting blots is fraud. "Proofer" or not.

We noticed this years ago when looking at -- IIRC -- ikaros antibodies. They were clearly faked. Lacking any sort of platform to gain attention we moved on to Abcam and our lab just sort of maintained a mental map of who not to purchase ANYTHING immuno- from.

The only reason I think biotech companies are not yet raising hell (and invoking the False Claims Act) is that Thermo Fisher's antibodies are already known to be notoriously bad, and everyone serious seems to have to validate everything themselves.

Exactly what is the "data" that's being shown here? Is it essentially some kind of marketing material showing "this sort of thing is what you should expect to see" or is it actually data or for compliance? If it's essentially marketing material or an instructional example that isn't meant to be representative it being magically clearer than real life doesn't seem like a great sin (unless it's being claimed it is representative). If it's something to be relied upon for compliance or as data to be used, that's pretty damming.

> "This image is supposed to demonstrate that the antibody being sold works as intended. It is labeled as “Advanced Verification” data on Thermo Fisher’s site"

(links to https://www.thermofisher.com/uk/en/home/life-science/antibod...)

I think it is technically marketing material, but if you have to fabricate your marketing material, that's not a good sign that the material is accurate. If I buy a car based on an advert where it shows the car going at 300 mph, and in real life it maxes out at 30mph, that's misleading advertising and something should be done about it.

Given that "at Thermo Fisher, a single vial containing a 0.1 mL aliquot of antibody solution typically costs 400 to 500 USD", you'd want to have accurate marketing material before buying it

It definitely isn't a good look but I'm not entirely sure where this lands on "the line drawing on my IKEA instruction manual doesn't look like the furniture" to "VW diesel emissions report" spectrum. I'd appreciate if any bioscientists in the audience could clear that up a bit.

From the article:

> This image is supposed to demonstrate that the antibody being sold works as intended. (…) Antibodies are near-ubiquitous but notoriously fickle laboratory reagents in biomedical research. For many applications, it is absolutely crucial that the antibodies that you use are selective (i.e., the antibody binds strongly to the target protein) and specific (i.e., the antibody binds to the protein of interest and little else).

Antibodies showing a different picture (Western blot) than what is expected can drastically change the interpretation of the results as well as the conclusion of a study, for example. It may also encourage scientific fraud by authors by forcing them to unknowingly/coincidentally make to a blot image the same (or similar) fraudulent modifications performed by the vendor.

Now I’m curious about how much of the blot photoshopping present in retracted papers can be attributed to these misleading verification data.

I would be more worried if the blotted area was different (the dark blob) - or if data in a datasheet (something like test specificity, level of detection, etc) was wrong

Now, if while preparing the images they needed to do some editorial choices (or it is well possible a person in the editorial group was told to 'enhance the images' but wasn't aware of the details) because of limitations in doing the experiment then this is probably not a big deal

Do you work in biology?

> would be more worried if the blotted area was different (the dark blob) - or if data in a datasheet (something like test specificity, level of detection, etc) was wrong

These images are provided on the datasheet and form the basis for the level of detection / specificity claims

> I would be more worried if the blotted area was different (the dark blob)

Or if more than one blob is present (i.e. blobs at different molecular weights) for a supposedly selective and specific antibody that should show exactly one blob on the blot.

> Now, if while preparing the images they needed to do some editorial choices

Editorial choices on raw scientific data are a big no-no.

> Editorial choices on raw scientific data are a big no-no.

I don't think you can find a picture in an article that hasn't been photoshopped in one way or another (which is mostly ok as long as it is not misleading)

Edit: TF's reply is interesting https://www.thermofisher.com/es/es/home/life-science/antibod...

Basically they say they are reviewing the images

Usually, journals require raw, unmodified data to be deposited as supplementary information.

Yes, as supplementary information

(Also journals are usually more rigorous than marketing material)

> Yes, as supplementary information

Still part of the article.

It is obvious that they edited the images to make blobs look shorter vertically. And in some cases, simply copy-pasted non-existing blobs.

This is 'used car salesman' levels of fraud on that scale. People rapidly acknowledge these antibodies work or they don't. There are websites with reviews of them. However in addition to getting ripped off for a few hundred bucks, these antibodies are generally produced by immunizing animals and by faking this data they are unnecessarily increasing the discomfort to these animals for a fraudulent reason. Look up the Santa Cruz antibody scandal for more of that.

Monoclonal antibodies can't be produced by animals.

Antibodies on this scale are not _produced_ by animals. The way these things work is an antibody that binds to a specific target will be _discovered_ by immunizing animals and then screening it's blood or lymph nodes for antibodies of interest. Once an antibody is found and commercialized, it is produced in large bioreactors of cell cultures engineered to produce large amounts of a particular protein (in this case the antibody).

Source: work as a software dev at a company developing antibody drugs

It's worse than this -- it's often very difficult (expensive, time consuming, etc) to generate the sample being run on the gel, the impact of using a bad or misrepresented antibody are significantly higher than the cost of the antibodies alone.

Images like this show how specifically the antibody binds to the antigen. Generally, the ideal is to have very specific binding. As such, this type of image (Western blot) would only have single bands in any vertical lane. Any other bands show that the antibody is binding to other molecules.

The evidence of painting out the background is likely someone cleaning up other bands, where the antibody has bound to something other than the intended target. So, they are making out the antibody is better than it actually is.

Copy-pasted bands could be evidence of attempting to make a weak band look stronger, or even adding a band where one didn't exist - potentially the entire blot is fabricated.

Either way, like someone else said, this is like fabricating parts of a data sheet.

It doesn't excuse it, but like someone else said, scientists would never just trust an antibody they bought. They'd do their own tests. Labs will also share notes amongst each other, along the lines of "that antibody is bad, and also strongly binds XYZ. You should try this other one instead".

(forgot to say I'm a former post-doc in Cell Biology)

The closest we all will ever get to getting meta meta answers on HN.

I'd treat this about the same as datasheets for mechanical or electrical parts.

When I buy an electronic component as a regular consumer I expect the datasheet "typical" values to be accurate 90% of the time. I can imagine larger industrial customers would really raise a stink if it's worse than that. However, any critical components in my circuit must be verified and "binned", and that's on me.

This is the thing. Yes, the marketing material is bad. But, no one in lab trusts an antibody just because of where you bought it. A new antibody always gets tested and validated before use.

That is to say, this looks bad for Thermo Fisher. But, that’s as far as the damage should go.

Why would you even generate fake pictures of this type? Don't you already have real ones? I mean, it's actually more work, unless you don't have the real ones.

I’m not going to defend Fisher here. It was a stupid thing for someone to do.

But unless you’re in the field, you won’t realize exactly how big ThermoFisher actually is. They are the major supplier of everything for molecular biology work. From freezers (the Thermo part) to plates and pipettes (Fisher) to enzymes and antibodies. In many ways they are like Amazon. They sell everything. Some of it from outside companies, but a good deal of sales are from in-house brands. They could use their position as a reseller to know which products sell the best and with the highest margins.

In a company of this size, it’s easy to have one group feel pressure and cheat on running the gels to confirm results. Particularly when the real results are ambiguous or dodgy. It’s not a good look, but I doubt it will put a dent in people from buying things (non-antibodies) from them.

Would it be the same idea as an x ray of a critically welded part?

It is intentional fabrication: it requires a lot more brain cycles to surreptitiously produce false data, it would literally be less work to Prepare the Western blots and scan them. Its the stuff they sell, so they have ready access to the products and would be much easier and cost-effective to simply perform (if it worked). Only if the product is known not to perform as specified is there a profit incentive for fabricating such "evidence".

It's more than just false advertising, it's criminal negligence wasting research attention, research time (repeating experiments to understand whats not working), naive nameplate quotations in the scientific literature also corrupts the scientific record (the author knows they are simply restating the nameplate specifications, but the reader may confuse it as a claim by the author).

Wondering if its sort of OK because it might just be marketing material, think of how the tobacco and other lobbies manipulate the scientific record. I mean technically it is marketing material... if one cynically views the scientific record as a poster wall where the highest bidder is allowed to plaster their spam all over the place.

Doing a western blot right takes a bit of practice and there are a couple failure modes you need to watch out for. Stuff like background "noise", smears, drifts can make it hard to get binary decision out of your experiment. E.g. antibodies are usually very very specific, but they can have impurities, unspecific bindings to other proteins etc which make interpretation harder. If they remove these from the advertised images you'll have a hard time comparing your own results to them. ESPECIALLY if they remove whole bands from the gel picture, which imho should be very much verboten.

Typically these catalogues have some numbers with regards to the antibodies binding affinity / impurities so you can have a general idea of what to expect, but having a clean image might mislead you into thinking that you did something wrong in your own setup. Seeing how wide spread it is, it's easy to imagine that their own lab is not run very "cleanly" and they have antibody contaminations in their gels, or issues with their own protocol that they're trying to edit out. Doubt that's the case, but it's really not a good look.

When you’re deciding which antibodies to buy you’ll look at these figures to get a sense of their quality. Antibodies aren’t perfect and might bind proteins unrelated to the one you want to study. Depending on your application, some off-target binding might be acceptable, but usually it’s not. Also, they might just be completely nonfunctional and bind nothing (perhaps due to missteps during manufacturing).

So what this fraud does is convince you to give these antibodies a chance when you otherwise wouldn’t have. You should validate them yourself and show they only bind your target before doing an experiment, but now you’re just wasting time and money evaluating something that’s guaranteed to fail.

Antibodies are notoriously unreliable, so you might have to give two or three vendors a try before you get one that works. Now I’m starting to wonder how much of that reputation is due to fraud and not just nature.

https://www.thermofisher.com/us/en/home/life-science/antibod...

> Moving forward, where an original image is not present or available, the Company will ensure that website users are informed that antibody images may have been optimized for presentation and clarity on the website.

wut. Bro if you don't have an original valiation image then the answer is not to say "oh we'll make sure we communicate that we're making up a random image" - it's to say you don't have the damn image. It's validation data wtf. It's not a pretty background image it's validation data if you don't have the data wtf are you "optimizing for presentation?" This faq is unreal - pure CYA except by someone who doesn't seem to know what they're trying to cover. If you've got cut and pasted/rotated bands that's just fake data. Not "optimized for presentation."

Yes labs should and usually do always validate new antibodies as well. It's a waste of time and taxpayer money for them to spend their time on bad antibodies they purchased based on fake validation data. And just fundamentally - don't make up validation data. If it's not there it's not there. What are you optimizing for presentation if there's no original!? What does that say about the rest of your process?

> 6. Did Thermo Fisher manipulate or fabricate antibody data?

> No.

Listen to these guys. What assholes.

And this is just fraud that was done with incompetence, so easily caught. How much is done competently?

Concerning but not really surprising. They offer about hundred thousand antibodies, a few hundred frauds is likely the tip of the iceberg.

> “Similar image” searches using Google Lens, Bing Images or DuckDuckGo betray hundreds more that we have yet to document

In my experience these would return any image of an antibody (edit) Western blot, not just the exactly matching background. Would be curious to hear others thoughts.

Without engaging in your point, small nitpick: These are images of Western blots[1], not PCR.

1: https://en.wikipedia.org/wiki/Western_blot

From the comments on the article by the author it looks more like 10% so far and they haven't systematically looked. That means ~10% if a probable floor of how much fraud there is.

Not according to the complete comment:

     More like 10%, but my search has not been systematic. I am mostly looking where I know I will find image issues based on image filenames and “Find Similar Images” searches.

They are clearly saying they think this is likely above average.

There have been efforts to standardize antibody reagent testing that are sorely underfunded/undervalued, https://ycharos.com/ (https://www.nature.com/articles/s41596-024-01095-8)

I have already forwarded this to a number of senior pathologists and industry leaders. People who are in charge of either developing these things or validating them for clinical use. We'll see what they say. But I suspect a great many of them will actually continue using the implicated products because they passed clinical validation at many institutions. TBD.

I have no idea about this catalogue, however, looking at the article and how the image manipulation has happened - it looks very much like "repro" work back in the day.

Anything that large companies published in/as magazines, etc, back in the 80/90s first went to a design company. Then to a repro company for the "finishing touches" to make it look nice. Faces were touched up, photo artifacts was removed, everything was to look neat and tidy.

This looks so much like that. I wouldn't be surprised if Thermo Fisher still ran everything that is to be published through a marketing/repro cycle, who has tampered with this without realizing what it looks like.

It'll be interesting to see if any actual data has been changed, or just the presentation of the data.

No marketing or design company would duplicate a band from another experiment (taking care to rotate it to make it look different nonetheless). Even in that unlikely scenario, Thermo Fisher is still responsible for the scientific data they publish.

The background painting could maybe be explained like this (depending on what was hidden), but I don't think the duplicated blobs have a good explanation, especially because some were rotated to try to hide the manipulation.

I disagree. If you look at the photos the painting used to make black blobs shorter. As I understand, black blob vertical position is the weight of a molecule, and they want to hide the fact that there are heavier or lighter molecules. So originally there was a long blob, and they made it look shorter.

In some of the experiments, the same random noise background exists with different black blobs superimposed at where you expect the correct value to be. Ie they took a fixed realistic-ish looking background, and drew in the 'correct' values

Its hard to argue that that isn't fraud as a result. It isn't touching up existing data, its fully fabricating data

Have the samples found so far, in general, been edited in a way to increase value or potential sales volume? Or are they just more pretty?

Some are just completely fabricated so it's hard to say if they have equivalent uglier images with real data...

More pretty, which would signify that the sample has less impurities -> better value

My most generous interpretation would be: the marketing/website team didn't get the pictures in time from the respective teams, so without much thinking they edited some. Like those print-on-demand t-shirt websites that don't have real models wearing the real shirts but crappy photoshop composites.

Someone call bobbybroccoli, they've got a new video to make :P

Unfortunately, the equipment and reagent vendors in the very unregulated life sciences/biomedical research world constitute a racket. Serial buyouts over the last 10 - 15 years have led to a ridiculous degree of consolidation -- the sort that wouldn't fly in a regulated industry, or even one in which regulators are paying any attention at all -- so it's now dominated by two players: Merck (through its MilliporeSigma arm), and Thermo Fisher. The existence of this cartel means that they can essentially get away with murder, both from a fraud angle (which is exactly what the Western blot manipulation is) and by fixing prices to whatever degree they please.

Also unfortunately, biomedical scientists are not known for their tendency to collaborate to face a mutual enemy (the mild pushback against the Elsevier/Springer Nature publishing cartel has come less from scientists and more from university systems whose libraries have to foot most of the bills). From their perspective, it's "What am I gonna do? Raise my own antibodies? Start blowing my own glassware?" So they grimace and bear it.

For reference, here's how their workflow for research antibodies goes (and it's been like this for decades):

1. Produce an antibody the research world needs. Do no QA, that's expensive and unnecessary.

2. Claim with usually no evidence (and apparently by forging evidence) that the antibody works in certain applications.

3. Let researchers buy the antibody and do your QA for you. Even if the antibody doesn't work, only a tiny percentage of buyers will go to the effort of getting a refund.

4. Profit. Keep selling the antibody even when the rare scientist with time on their hands demonstrates beyond doubt that your antibody clearly doesn't work.

5. When sales start drying up because enough people are catching up to the scam, discontinue the antibody. Give no explanation.

6. Repeat from 1.

Holy shmoly... I'm a biologist who has used Thermo antibodies before, and this is seriously disappointing to see.

It's like 90% biomedicine papers are like this.

True.

Reminds me a lot of the Schon scandal. TLDR is (now-obviously-a-fraud) generational physicist kept publishing breathtaking work about semiconductors, was caught because two of his error distributions were identical

why isn't their stock tanking

Quite shameful of Thermo Fisher. Therapies are based on accuracy. Did they damage people by lying to them?

Also, how many other scientists just bought into that and used this for their own "analysis"?

Western Blot is a hideous technique

What would happen if I drank these antibodies?

Probably not very much, same as if you ate agar, grass, a pencil eraser (rubber), or candle wax. But why would you?

You'd become a new character in the MCU: ImmuneMan

An antibody is a protein. The acid in your stomach will break down the molecules into their constituent amino acids, which would be absorbed into your bloodstream. The chance that the antibody would have any biological effect before being broken down is slim. Still there are other reasons for not drinking it such as whatever else is in the solution such as preservatives etc.

What if I injected it?

Your insurance will probably deny it, and you'll get a bill. ;-)

My guess is possibly not much, but it could vary greatly from one antibody to another. If the antibody isn't specific to any human protein, then it'll probably end up in your pee eventually. On the other hand, there are therapeutic proteins that are given by injection, so I could hardly suggest that it's intrinsically safe. And antibodies made for medicinal research stand a good chance of existing because of possible human interaction.

Apparently, the authors of the original paper haven't yet learned to use AI. Their successors, however, will almost certainly do a better job of it.

What paper? This is about several hundred datasheets.

Reply on news.ycombinator.com