Investigator: “Dr Latchman, as senior author, should bear the weight of responsibility as there was no other author in common”

Update 30.08.2019: One of Latchman investigators speaks out. John Hardy, award-winning geneticist, UCL professor for neurology and Fellow of Royal Society, contacted me yesterday with this statement:

Dear Dr Schneider
I was referred by a colleague to your article on the Latchman affair (https://forbetterscience.com/2019/07/08/the-return-of-david-latchman-show/) which is largely accurate. I was the chair of the first committee investigating the allegations and, when appointed, fully expected them to be baseless. However, to my horror and that of my fellow committee members, nearly all of the ones we could check were correct (see pubpeer https://pubpeer.com/search?q=+David+S+Latchman). Our report was very clear that Dr Latchman, as senior author, should bear the weight of the responsibility especially as there was no other author in common to the papers. Our report was not at all a cover up although it was watered down in response to objections from Dr Latchmans expensive lawyers. The chief weakness was that there was no direct evidence that Dr Latchman himself had ordered the fraudulent image manipulation endemic to many of his papers. When our report was released we were horrified that his current employer Birkbeck University put out a press release saying he had been cleared himself of fraud and objected internally to the way this release misrepresented our report.

There were many papers we did not investigate for lack of time and other allegations of fraud were made. A second committee, completely independent of ours (I did not know it’s members until after it reported) investigated the second tranche of papers and reached exactly the same conclusion with exactly the same outcome with again, a very misleading response from the Birkbeck press office. That committees members, like mine were extremely unhappy with this.

The fundamental problem is that unless there is documentation that a PI has instructed the postdoc or student to fabricate data the burden of proof of guilt cannot be achieved.

I would additionally note that the only types of fraud we can see are image manipulation and the use of the same images with different legends in disparate manuscripts. As a glance at pubpeer will show there are multiple examples of this in his work. The idea that all other papers from the Latchman lab are reliable would seem to me to be very unlikely.

I would also suggest this overall constitutes financial fraud. since I presume these data either directly or indirectly were used to support promotions and grant applications.

In my mind last/senior author position means one is taking responsibility for the integrity of the work. Dr Latchman and Birkbeck apparently do not agree. And as Master at Birkbeck he will be the final arbiter of any other allegations of fraud at this once proud institution.  He also sits, apparently, on National Ethics Committees (https://en.wikipedia.org/wiki/David_Latchman). I think this is an unfortunate outcome.

John Hardy PhD FRS

---

Smut Clyde’s and my reporting in the affair of David Latchman, heir of a billionaire philanthropist, geneticist, UCL professor and Master of Birkbeck, apparently had an effect. The media, which almost accepted that Latchman was innocent of the research misconduct committed by his more junior collaborators, pressured UCL to release the investigative report under the Freedom of Information Act (FOIA). Turns out, in each investigation Latchman has been found guilty of research misconduct by being “reckless”.

Same FOIA request from my side was denied by UCL a year ago, to protect Latchman’s privacy. But then again, UCL rejects or simply ignores all my FOIA requests, on principle. As I was told, in December 2018 UCL was “minded” to release the documents to proper journalists under FOIA, in February 2019 it was decided to comply with the requests. UCL then released the files (not to me, obviously), after a thorough cleansing. The censored reports were provided to Buzzfeed and Guardian, but only the former also made the full documents available online.

Lawyers, lawyers everywhere

The annual investigations of Latchman between 2014 and 2017 happened because of persistent email notifications by the data integrity sleuth Clare Francis, who also produced much of the PubPeer evidence, which now stands at 68 papers. Every single investigation report found the Master of Birkbeck guilty of research misconduct by recklessness, due to his impressively abysmal supervision and his obsessive desire to be main coauthor on every paper leaving his lab. Yet in the media so far, Latchman was rather a victim than the perpetrator, see these earlier news stories here or here. The newly released investigative reports show something else. Buzzfeed writes now, tongue-in-cheek:

“The fact that UCL took five years to investigate the allegations, and then issued no disciplinary sanctions against anyone despite widespread evidence of scientific fraud, will add to lawmakers’ concerns that Britain’s universities are failing to adequately police research misconduct.

UCL declined to comment on why it did not take any disciplinary action against Latchman. “We will not be adding anything further,” UCL spokesperson Kirsty Walker told BuzzFeed News by email.

Shimon Cohen, a reputation management specialist working for Latchman, told BuzzFeed News by email that Latchman rejected the assertions made in the investigation reports.

“The nature of the manipulation identified in the UCL investigation was such that any fraud would only be apparent to a reviewer who was actively looking for such deception,” Cohen wrote. “To subject all research to this disproportionate scrutiny is not reasonable, to then make the assertion that to overlook such deception is ‘inattentive’ or ‘reckless’ is unjust.”

See, checking someone’s paper for fabricated data is unjust. As opposed to publishing fabricated data, presumably. In any case, a reputation manager is only engaged by those in deep trouble, in Latchman’s case combined with very, very deep, in fact scarily bottomless pockets. After all, Latchman engaged one of the most expensive London law firms, Mishcon de Reya, to help UCL with the investigation of his papers. This is probably why neither Guardian nor Buzzfeed dared to even indirectly mention the money issue. All these millions of pound sterling which Latchman’s Wohl Foundation (named after his late billionaire uncle Maurice Wohl) keeps giving to Birkbeck (read here). Sources informed me that Birkbeck is so tightly controlled by UCL in all but the name, that the Wohl Foundation money given to Birkbeck is very likely to serve UCL also. This is why this UCL statement to me bears little meaning:

“UCL has had no donations from the Wohl Foundation in the last 10 years.”

As those sources also explained to me, the investigative reports released by UCL (in a heavily censored form) are not entirely original. The final reports compiled by the UCL investigators were namely processed and reformulated by Latchman’s lawyers, and then UCL bureaucrats had another go at those. For example, the report’s authors were only able to scrutinise 8 papers, due to time and workload constraints, and found all of these 8 to be fraudulent. Other suspect papers reported by Clare Francis were left unaddressed, which is not exactly same as exonerated. Yet in the final version, these other papers were suddenly decreed as perfectly fine, and the reason why they were not scrutinised became the allegedly unsubstantiated evidence.

Panel of investigators: We didn't have time to look into 20 of the papers in the list but the eight we checked were full of made-up stuff.

UCL: The panel found nothing wrong in 20 papers! Latchman was completely vindicated! pic.twitter.com/xk2FspaRjN

— Smut Clyde, X-Ray Haruspex (@SmutClyde) July 3, 2019

Even now, the released UCL reports censor which papers were found fraudulent, a ridiculous attitude which was explicitly overruled by the Information Commissioner in a different case – see my reporting on Irina Stancheva and Edinburgh affair, where the only part of the report I was eventually given was the list of publications found manipulated. Of course the scientific community has the right to know which papers are not reliable, especially in medical research there is also patient interest. But the universities tend to see it otherwise, UCL certainly does.

This is why Latchman so far retracted only 6 papers (and corrected 2), and not 12, or 25, or even 50: exactly because the list of fraudulent literature is kept secret by UCL even now. The authors only retracted what they absolutely had to, for whatever reasons (eg, when demanded by the journals), and not what the investigators asked to retract. Lawyers are apparently heavily involved also now, research fraud is a booming legal industry these days, as my own court sentences prove.

Observe the change. An 2018 article in The Telegraph named two former UCL researchers fingered for fraud in Latchman affair:

“The panel found that two scientists – Dr Anastasis Stephanou and Dr Tiziano Scarabelli – were guilty of misconduct by manipulating images in seven published research papers. Prof Latchman, former dean of the Institute, is cited as an author on all seven papers.”

Only Anastasis Stephanou, who works in Nicosia and still claims to hold a UCL assistant professorship (from which he has long resigned from, according to UCL), has been named now. The European University Cyprus professor and “author for over 150 peer-reviewed articles” was quoted in the Guardian admitting his responsibility for image manipulation, while in Buzzfeed Stephanou denied all responsibility for any fraud in the papers where he was last author. And as for the other now suddenly unnamed perpetrator, the cardiologist Tiziano Scarabelli... That one was recently sacked by his Virginia Commonwealth University in USA (certainly not for data manipulation, VCU cannot care less about that). The native Italian now sues the university and some colleagues for retaliating against his alleged activities to protect patients:

“according to a timeline laid out in legal documents, sexual harassment charges were raised against Scarabelli, and he was soon suspended from his position as the health system engaged in an investigation.

Scarabelli was cleared of the charges in February 2018. The following month, he was told his contracts with MCV Associated Physicians and VCU Health System would both terminate before the end of the year and would not be renewed.”

The lawsuit is set at $15 million, no wonder Buzzfeed and Guardian apparently decided to accommodate Scarabelli’s lawyers and never mention him.

Actually only half of the 67 Latchman papers discussed on PubPeer have Stephanou and/or Scarabelli as coauthor. But this is exactly why these other remaining 34 papers were never considered for investigation by UCL. Whom to blame for that utterly Photoshop-fabricated gel in Shulman et al 2002?

Or this creative recycling of gel bands, in Ensor et al 2001 and Ensor et al 2003? If everyone was free to fake data any way they liked it in Latchman lab, what kind of scientist is the big boss then?

Very reckless one for sure. But now, to the reports published by Buzzfeed. They are prefaced by a UCL statement:

“Several actions were taken responding to recommendations in these reports including:
1. A formal disciplinary process was initiated against a member of Professor Latchman’s research group. The individual [Stephanou, -LS] resigned and left UCL before these proceedings could be completed.
2. Two separate disciplinary processes into Professor Latchman found that there was no deliberate intention to commit misconduct in research and, therefore, there were insufficient grounds for dismissal.
3. Authors and editors of the relevant papers, including Professor Latchman, were directed to correct the record of research where research misconduct was identified – resulting in retractions and clarifications in the relevant journals.
Professor Latchman ceased to supervise research.”

Poor Latchman now has to supervise research only indirectly, in his position as university rector of Birkbeck, where he receives a yearly salary of GBP 400k, one of the highest in UK. One can hope the Wohl Foundation will also continue supporting Birkbeck and thus also UCL, despite all the ongoing kerfuffle.

The Latchman reports, censored

The first report of the investigation from 2014 addressed 28 papers. Please appreciate Latchman’s explanations, highlights mine:

“During interview, Professor Latchman also made the following points in mitigation:

• Professor Latchman explained that he considered the tone of the emails sent by the Complainant to be vitriolic towards him personally and that he did not accept that he was a ‘fraudster’ as had been stated within one of the Complainant’s emails.
• Professor Latchman noted that the number of articles covered by the allegations represented a small part of his total research output produced during his academic career to date (some 250 papers).
• Moreover, some of the articles that had been examined dated back to 1997 and it was likely that there had been some 25 breakthrough papers on the research concerned subsequently.
• It was noted that the data image technology and software now available was far more advanced than that available at the time that the older articles covered by the allegations were produced.
• Professor Latchman was made Head of Department of Molecular Pathology at UCL in 1991, and moved to to become Director in 1999, and as part of that appointment he secured a lectureship post [censored] Professor Latchman was keen to develop the research careers of his colleagues and started to use corresponding author status on work produced by [censored, presumably Scarabelli and Stephanou]. He subsequently became last author on their work so as not to hinder their academic development. However, Professor Latchman noted that still included his name on publications, although his role as last author was advisory given the work concerned was developed independently. Professor Latchman accepted that whilst it might be mistaken for him to continue to be named as last author, he did so to help [censored, presumably Scarabelli and Stephanou] in their applications for grant funding and promotion.
• Professor Latchman did not personally prepare any of the data in the articles covered by the allegations, nor did he directly supervise all researchers in his Lab.
• Professor Latchman remains as Head of the Medical Molecular Biology Unit at
  ICH, located at the Rayne institute at UCL, as having Head status allowed him
  to sign off grant applications by researchers in his Lab. He was also [censored, presumably Scarabelli and Stephanou] line manager.

As indicated in [his] written response and at interview, [censored, presumably Stephanou] explained that [he] no longer had any primary data prior to 2010. [Censored] stated that this had significantly reduced storage space, and that some of the experiments were carried out by former researchers in the Group who had taken some of the data with them upon their departure to complete their part of the manuscript. [censored] was aware that primary data should be kept for 10 years in line with UCL policy and accepted that he should have kept the primary data for longer and would do so now onwards.
Professor Latchman noted that he had spoken to [censored, presumably Stephanou] in the past about keeping primary data, as this was an area of concern. [censored] added that had published over [xx] peer-reviewed manuscripts and accepted responsibility as senior author on those papers where errors had been found by the Panel.”

To sum up: The Master of Birkbeck says that even if (by now) a large part of his research output was found manipulated it doesn’t matter because there are more papers which the sleuths have missed. Especially if he feels there were some “breakthroughs” among those somewhere. In this regard, Clare Francis explained to me how they analysed Latchman’s research:

“I looked at his output since late 1990s. Before then the image quality is poor. According to Pubmed he has 483 publications (6 of them retractions). 477. Many papers are in obscure journals/not freely available. Estimate I looked at 120 papers. 30 dodgy ones. There are about 20 papers, mostly 2000s without any primary data.”

This was the reason why merely ~70 papers of Latchman’s were found to contain shady data, and the rest not (yet). Nothing to be proud of. The Master of Birkbeck’s defenses make no sense. How can he seriously say that the Photoshop software was not as advanced in those days as it is now, so it is not fair to blame the his UCL lab for the hack jobs like this, in the papers by Mr and Mrs Scarabelli, Chen-Scarabelli et al 2005 and Scarabelli et al 2005:

The investigators indeed could never prove Latchman’s direct involvement into data manipulation, as I learned. But the indirect involvement, starting with Latchman’s appalling attitude to research integrity, was enough to find him guilty of misconduct (highlights mine):

“While the Screening Panel noted that it was not its role to determine whether the allegation against the Respondents was proven or unproven, it was nevertheless able to conclude on the basis of evidence presented to it that prima facie evidence of misconduct in research of fraud as defined within Annex 2 of the procedure as ‘deliberate deception (which may include the invention or fabrication of data) or other misuse of research funds or research equipment’ had occurred in relation to eight of the articles covered by the allegations. In light of the seriousness of the matter, the Panel considered that the matter should be referred directly to relevant disciplinary process or other internal process and bypass the Format investigation stage given the Respondents had accepted at interview that the articles be retracted.

The Panel would emphasise that it considers that a paper is only as good as the data in it. Therefore statements to the effect the data were in error, but the overall message of the paper was still OK, are not acceptable. If a single Figure in a paper can be shown, beyond reasonable doubt, to have been manipulated, the Panel recommends retraction rather than correction since the conclusions cannot be deemed to be supported by the data. The Panel would also advise that retraction was an honourable action to take, rather than being shameful. However, any retraction statement should be carefully worded to explain the precise reason for retraction.”

8 papers were set for retraction then, UCL refuses to say which. What we can check is that only 6 papers were retracted so far (2 were corrected). About the rest, we read this in the final, lawyer-processed report:

“The Panel also considered the remaining allegations made by [Clare Francis] on the other 14 articles and determined that there was no clear-cut prima facie evidence of research misconduct that needed to be pursued further.”

Because Clare Francis kept bombarding UCL with emails, there was a second investigation, in 2016 where 32 papers were analysed. This time, the investigators were external, probably recruited as per usual specifically on their trustworthiness to produce the verdict UCL would agree with:

“Having reviewed the allegations and the evidence available, the Panel took the view that, in the case of the following 22 papers, misconduct in research had not occurred and no further detailed investigation or discussion with the Respondents was required. This was either because the allegation was mistaken or unclear; because there was no evidence to support the assertion of misconduct; or because the allegation could not be addressed without access to original data, which were no longer available.”

Now that is clever. If original data becomes unavailable, there can be no investigation, no matter how fake the figures appear to be. In other cases, Latchman and his partners did get into trouble:

“In the publications considered by the Panel, three main categories of problem were identified:

i) Splicing of images of western blots, i.e. combining individual tracks from the get electrophoresis image from different western blots into the same image;

ii) Reuse of images of western blots and micrographs; Duplication or ‘cloning’ of segment of micrographs both within and between individual images.

Against the background of the guidance in the Procedure that standards by which allegations of misconduct in research should be judged should be those prevailing in the country in question and at the date that the behaviour under investigation took place’ [Annex 2, para /] the panel concluded that the first category, i.e. splicing of western blot images, was not uncommon practice at the time of the publication of these articles, and was not then regarded as unsatisfactory, although it was not considered best practice. The Panel therefore decided that the splicing of images of western blots did not amount to misconduct in research.”

Now gel splicing and “library loading” controls is the hill I am prepared to be called a rapist on (as it happened in BethAnn McLaughlin case). Unlike it is repeatedly incantated, it was never acceptable to rig a western blot. This gel in McCormick et al 2006 is utterly untrustworthy, because the loading control and the phospho-protein blots are spliced differentially, meaning they come from different gels which originally showed who knows what:

Granted, other evidence found in papers from Latchman lab suggested much worse practices:

“The second category, i.e. re-use of images of western blots and micrographs, could not be viewed in the same way. White the Panel considered that an occasional instance of such re-use might arise through simple error, for example by mislabelling an electronic file, there were numerous instances of the re-use of such images, and in view of this the Panel concluded that, on the balance of probabilities, in the instances identified above these manipulations had been carried out intentionally, and accordingly amounted to misconduct in research.

The third category of problem, i.e. the manipulation of individual images, was regarded by the Panel as the most serious category of research misconduct. This could not have arisen by accident. There were numerous instances of image manipulation identified by the Panel, for example, cloning of parts of the image”

Again, UCL censored the list of the papers, but let me give you a colourful example, an artwork in Scarabelli et al 2001:

No, that paper in Circulation is not retracted, it will not even be corrected. For all we know, it probably was never investigated. Neither will anything happen to Canalaris et al 2003, where this collage appeared:

The only way to get through all that was UCL’s literal pig-headedness in face of overwhelming evidence. Throw out the worst papers from investigation for whatever reasons, refuse to admit new evidence or evidence where raw data mysteriously disappeared, declare gel splicing an accepted practice, intimidate investigators by handing out to Latchman’s lawyers their email exchanges, twist the final report’s wording, control the press, and in the end: the Master of Birkbeck has no case to answer.

“The Panel did not consider that Professor Latchman had any intention to commit misconduct in research, or that he had any knowledge of the misconduct in research identified in the investigation at the time. The Panel noted that the referees of the publications, and the readers of [censored] had not identified any problems at the time.

In discussions with the Panel, Professor Latchman volunteered information consistent with him having been insufficiently attentive in the management of the laboratory; he explained that this had occurred because the group leaders had become more independent over time as they secured their own research grants.

The Panel concluded that Professor Latchman had been insufficiently attentive in the management of the laboratory and in respect of the group’s publications. His inattentiveness to the detail of the research going on in the laboratory over a number of years, and to its reporting in the papers published, had allowed the research misconduct by others to continue. The Panel concluded that, although he had no intention to commit research misconduct, his failure to manage the laboratory appropriately, together with his involvement as senior author or co-author on the publications where misconduct was identified, amounted to recklessness in the conduct of an aspect of a research project, which facilitated the misconduct in research identified in this investigation.

The allegations of research misconduct in respect of Professor are upheld in that whilst he did not intentionally commit the misconduct in research identified in this investigation, his recklessness in the conduct of his laboratory and his involvement as author on many publications facilitated that misconduct.”

The third report from 2017 addressed 7 more papers, its findings are very similar to the previous ones.

“The Panel concluded that, in respect of this paper, misconduct in research had occurred. The nature of the research misconduct identified by the Panel in respect of the papers referred to at paragraph [12] is that of fraud (as defined in Annex 2, paragraph 7 of the Procedure).”

Latchman was once again found guilty of misconduct by recklessness. But because UCL is not saying which papers were set for retraction, there won’t be any more than those 6 retractions, unless journals develop their own ethics standards.

---