House of Mind

After weeks of worrying about how the mandatory across-the-board 2013 budget cuts known as the sequester would play out at the National Institutes of Health (NIH), the biomedical research community now has final figures. The bottom line is as grim as expected: The agency’s overall budget will fall by $1.55 billion compared to 2012, to $29.15 billion, a cut of about 5%, according to an NIH notice today. That is essentially what NIH predicted as part of the 5.1% sequestration.

As a result, NIH expects to fund 8283 new and competing research grants this year, a drop of 703, according to this table. That number firms up the “hundreds fewer” awards that NIH officials warned of earlier this year. Including ongoing (already awarded) grants that are ending, the total number of research grants will drop by 1357 to 34,902 awards. The decline “reflects the fact that NIH’s budget is being shrunk due to the new budget and political reality, which is bad news for researchers and the patients they are trying to help,” says Tony Mazzaschi of the Association of American Medical Colleges in Washington, D.C.

NIH will try to keep the size of the average award consistent with 2012; it will not award inflationary increases for future years. The agency also expects to trim continuing grants. Grants that were cut up to 10% earlier this year because of budget uncertainty “may be partially restored,” but probably not to the original commitment level, NIH’s notice says.

To be honest, when I read this article my heart dropped a little because it highlights one of the harsh realities that people in science would rather not think/talk about. I’ve always felt it was a privilege to be able to do science with federal funds but this is still disappointing. I can’t help but think how much harder getting a PhD and a postdoc is going to be :( However, I understand that sometimes you just gotta do what you gotta do and cut corners when and where you have to.

Science is not for the faint of heart. You’ve been warned. 

Replication of empirical findings is the cornerstone of science. To me, reproducibility of the data is actually a qualifier for an experiment/project to be considered “science”. However, multiple groups are finding something that makes me and many other members of the scientific community REALLY uncomfortable- failure to replicate experimental results published in top tier journals such as Nature. 

The whistle blowing has primarily been done by pharmaceutical companies like Bayer or Amgen that scour published literature for promising leads. These companies search the literature for novel models, mechanisms, and tools published by top labs (usually because of scientific reputation) and try to build on their work. In many cases, this results in a dead end, as an alarmingly high number of publications are not reproducible and result in the loss of thousands or even millions of dollars. And to think that the money could be spent on making real progress towards drug development…

The article above mentions some of the possible factors contributing to this situation. First, experimental methods rarely control for all possible variables. Limited funding also  puts pressure on the PIs push out “good” results for publication before conducting replication experiments and many times reviewers don’t ask for replication experiments. There is also the possibility that some groups provide the minimum amount of information necessary for publication because they are competing with other labs and are trying to safeguard their experimental procedures. Then we have experimenter bias: they see what they want to see and make a case for it. Finally, failed replications are seldom published, which means that if a group can’t replicate the results they will probably keep it to themselves instead of publishing discrepant results. 

Clearly, “failure to replicate” is a time-consuming, frustrating and costly process. Elizabeth Iorns, co-founder/CEO of the Science Exchange and the author of the linked article, has partnered up to create the Reproducibility Initiative-a new program to identify and reward high quality reproducible research. Here,  results can be submitted for independent validation to more than 1000 expert experimental providers. Validations are conducted blind, on a fee-for-service basis. This is an interesting approach, but will it really solve anything? For example, the initiative is on a voluntary basis (researchers submit work they want to see replicated) and scientists would have to pay to have their results replicated, which seems like an additional cost that not many scientists would be willing to add to their expenses. Moreover, who will these “replication experimenters” be? I can think of more than one situation in which it may result in a conflict of interest. Will they be made publicly known, do you get assigned a group or do can you pick who gets to replicate your work? Last but not least, what if the results are not reproducible? Will there be a way of making this publicly known or will it just be swept under the rug like nowadays?