Getting what you paid for in quality control? Cell lines exemplify a more general challenge.
Andrew Moore Editor-in-Chief Recently introduced requirements for cell-line authentification in some journals are a very positive development, but might this distract attention from a more general problem that should also be monitored constantly, earlier in the research chain? A summary of decades of individual researchers finding certain cancer cell lines not to be what others thought they were 1 should make us reflect anew on quality control in research. When you buy a new washing machine, you hope that it will wash your clothes well; but you also hope that it won't mangle them or leak water and damage your house – and, you hope it won't damage your health, but more of that later. For these reason, if you can afford it, you will want a machine from a well-established, respected manufacturer: one that has undergone sufficient quality control in its design and build to assure you of good performance. The said manufacturer also has an interest in applying quality control: it defends his pricing model, strengthens brand penetration, and avoids costly legal cases and compensation. On the last point, the consumer has a legal right to take the manufacturer to task over certain failings because the purchase secures contractual obligations. Basically, in the world of manufacturing and consumption of household appliances there is a self-reinforcing system that rewards each side for choosing quality control. In publicly funded research, the research funder is the proximal paying “customer”, so to speak. But the research funder has the responsibility of representing the population – distal customer – whose taxes have paid for the research and – in the case of Open Access – its publication. The proximal consumers are, on the whole, other scientists at subscribing institutions; and – in the case of Open Access – anyone with an Internet connection and a computer. Regardless of the business model, very few of these individual consumers personally pay for the product, and hence a part of the quality control culture and responsibility that surrounds banal items such as washing machines is diffused into a kind of no-man's land of institutional, community and corporate responsibility, and often inaction. Part of the concept of standardized quality control is extremely hard to apply in science because we are dealing with the production of bespoke items: each research publication is unique (well, overwhelmingly). Each has to be assessed by its own specialist quality controllers. And whereas QA experts in industry are trained and accredited – i.e. quality control of the quality controllers – there is no such system in science. But another aspect of research quality control refers to research-generated materials used in the work that precedes publication(s): materials that are shared and used within a community. This aspect of research is much further removed from independent scrutiny, and it is individual scientists that are at stake here. An individual scientist – and her/his research group – is the basic entity of consumption. That individual has not explicitly paid for a given research article that has contributed to steering her/his research, or a given shared cell line, and I have the feeling that a particular default attitude applies: no financial contract, no obligation, and certainly no guarantee. And yet the receiving researcher's work could be damaged by the faulty product of another – be it a publication or a cell line. The damaged party is, in a sense, represented – as a tax-paying member of the population – by research funders. By extension, the research funders bear some of the responsibility to institute quality control of research before it reaches the point of publication; that is currently done at the point of grant application assessment, but arguably it should apply at intervals thereafter as well. This will be more or less challenging, according to the particular research, but one level at which it should be possible is the quality control of biological resources: after all, even in the production of a commonplace washing machine, respected manufacturers regularly quality control their materials. As noted in 1 in the case of false-identity cancer cell lines it is genuinely very hard to identify and extirpate them, though this particular quality control problem has potential consequences for the treatment and prospects of individual cancer patients. But a poignant message here is also that awareness and application of a technical innovation of enormous power was not nearly as strong as it should have been. Quality control of research-generated materials is not a mere technical adjunct to research: it is a science and technology in its own right, and its developments need to be followed and applied with as much diligence as that applied to research. Research funders can help towards this aim. Andrew Moore Editor-in-Chief
Read more
1
