Writing a K99 was one of the most challenging and ultimately rewarding endeavours of my academic career. Since the awarding I’ve had a number of requests for advice, and while I obviously do not have the definitive recipe for success, I thought it might be useful to post a description of my experience.
In writing my grant, I relied upon a number of excellent resources -- I highly recommend beginning with these, and so I have included links at the end of this article. These resources provide definitive advice for writing the grant, which I followed and I think does not need to be reiterated. Instead I will focus on the strategic planning that I undertook before I did any actual writing -- something that I found to be missing from the resources that I found.
I’ll be blunt. This grant is a ton of work and not something that you should take lightly, or try to complete at the last minute. If the deadline is less than a month (or two) away, do yourself (and anyone whom you are imposing upon for letters/advice/assistance) a favor and reconsider. This is a highly competitive funding mechanism and success requires planning, organization, and coordination with your mentors, references, and institutional grants officers. Here, I outline the step-wise approach that I took to planning my grant; in each section is my advice and the underlying logic, followed by my experience.
Determine your eligibility window
Count backwards: you ideally need 3-6 months to put together a competitive application and the recent shortening of the eligibility period to four years of postdoctoral training presents additional challenges. You should also give yourself the opportunity to apply twice: getting your grant to the review stage is a very reasonable goal for your first submission, and you can then address criticisms and improve the grant for the second submission. Altogether this means submitting at least eight months before your eligibility expires, potentially more depending on how your eligibility falls with respect to the three submission deadlines (in February, July, and October). It also means you may need to start your first application nearly a year before you complete your fourth year.
I strongly recommend submitting twice. I know a lot of applicants who deferred until they had a paper, failed to get an award, and were then ineligible to resubmit. The value of a submitting twice far outweighs the costs: it is extremely unlikely you will be awarded a K99 without postdoctoral publications, but the feedback you receive on the application itself is invaluable. In that first submission, outline the publications you expect to have, but don’t oversell it; you may be penalized for failing to deliver in your second application. In the resubmission, you can then address the criticisms, list the papers you published in the meantime, and demonstrate that you deliver on your promises.
I submitted under the previous five-year limit, with my eligibility set to expire in Sept 2013. I initially submitted in Oct 2012, hoping only to get reviewed -- scored a 44, which was quite disappointing. I published well in my PhD: a first author and co-first author paper in high impact journals as well as a middle authorship on eight other papers, many of them in high impact journals such as Cell and Science. However, I had nothing notable from my postdoc at the time of the first submission: 2nd author on a review. Due to the length of the review process I missed the next February deadline, but was able to submit a revised application for the July deadline. This resubmission was successful, scoring a 17. By then I had delivered on my promise to prepare and submit two first-author papers (PLoS Genetics, accepted; PLoS ONE, in revision), and I had a bonus publication (2nd author Nature Methods). These weren’t classic glamor publications, but they demonstrated productivity, and may have instilled confidence that the other papers in preparation are likely to be published as well.
Understand the submission process
Familiarize yourself with the process of NIH peer review, if you are not already. The NIH website provides a good overview of how grant review works in general, and there’s a great video of a simulated NIH peer review panel (starring my mentor, Keith Yamamoto). The video depicts an R01 review, but much of it is applicable for what happens in a K99 panel. However, there are some important differences. First, since the K99 is a training grant you will be judged on five equally weighted categories: i) you as a candidate; ii) your training environment (ie. your home institute); iii) your mentors; iv) your research plan; and v) your career development training plan. I cannot stress this enough: the candidate background and career plan documents require just as much consideration and time as the research plan itself. The K99 grant is an investment in you as a researcher, so do not make the mistake of thinking that a good scientific idea should speak for itself.
The description below is based on the process that was in place for the 2013/2014 cycles. After you submit your grant, assuming it passes the basic formatting requirements, it will be sent to up to five people for in-depth review. These reviewers will judge your grant on the five areas listed above and they will assign numerical scores for each (1 is the best, 9 is the worst); they will also assign an overall impact score. These scores will determine whether your grant makes it to the panel for full discussion. This brings us to the second difference: the grant is not sent to a specialist study section, but rather to a panel of 20 members convened to review K99s. My understanding is that many or all of your reviewers will be on this panel (depending on if mail-in reviewers were used). This means that you need to write a grant that is both clear and appealing to a broad scientific audience, not a group of experts in your research area. Any panel member with a conflict of interest will be recused. The remaining panel will discuss the earlier reviews, and each will assign their own overall impact score. These impact scores will be averaged and multiplied by 10; now 10 is the best score, 90 is the worst (don’t try to understand it, just go with it). You will receive this overall score, along with the detailed reviews, the individual reviewer scores, and a summary statement from the panel; you will not receive the impact scores from the individual panel members. You need a score in the low 20s to be competitive for most institutions.
Both of my submissions were evaluated by five reviewers. In my first submission, the average impact scores from the reviewers did not match the final score: my reviewer scores (assigned by the reviewers before the panel convened) averaged at 17.5, but my final score was 44. Even when I averaged the worst scores from each reviewer for each category, I would have had a 30. So obviously the panel was far less enthusiastic than the reviewers, so where did it go wrong? From the summary statement I was able to glean that the panel had concerns about my ability to successfully execute some of the analyses, and these concerns were first raised in the panel discussion by an expert in bioinformatics. The panel also discussed my lack of postdoctoral publications and an over-ambitious proposal. Neither of these were significant issues for the reviewers, but came to light in the process of panel discussion, resulting in a much lower score than the individual reviews indicated. All points were very fair criticisms that I was able to address in my resubmission.
Come up with a rough project outline and develop Specific Aims
Your project should build on your previous work, but must be a clear departure from your work with your postdoctoral mentor. Take extra time and energy to clarify your research question, devising a research program that is broad enough to warrant long-term investment in you. This is not a postdoctoral fellowship, it is a full five-year research program that is more akin to an R01, and so it might require some adjustment of your perspective (throw away those myopic spectacles and opt for a wide-angle lens). Give yourself plenty of time to develop this outline -- you might be surprised how much a seed of an idea can develop as you discuss it with peers/colleagues/friends/your ever-suffering spouse.
Your Specific Aims are a key part of the submission. This one page document should include background, hypothesis, short and long-term goals; it’s essentially your grant distilled down to a single page through which a reader can rapidly get a sense of your scientific problem and how you plan to approach it, usually through 2-4 aims. Each Aim is a collection of experiments that cumulatively address an aspect of your larger question. Be sure to get multiple examples from colleagues. Pass your specific aims around to your mentors, colleagues (your ever-suffering spouse) to get critical feedback.
I had a background in C. elegans as a model organism for DNA repair and metazoan development, and had expertise in genetics, molecular biology, and biochemistry. The conceptual departure from my graduate (and current postdoctoral) work was a shift from modeling human biology to modeling the biology of a parasitic nematode (Brugia malayi). To successfully complete my proposed research program, I argued that I needed additional training in systems biology approaches, analysis of large datasets, and working with the parasite. I took 2-4 weeks to develop the project before I even started writing.
I used an iterative approach, coming up with a rough outline, developing specific aims on it, passing it to colleagues, and then revising the aims in response to their criticisms. This process took me almost three months, from May 2012 to August 2012. During this time, I did no other work on the grant, but did maintain my bench work.
A piece of new advice, added since the original post is to find 1-2 "designated ass-kickers". These are members of your support network who will provide unflinching, brutally honest advice on your specific aims, grant, presentations, etc. If they think that an idea is flawed, they will tell you in no uncertain terms. It's better to get this level of criticism very early in the development of the grant, rather than having it come from a reviewer, prospective employer, etc. This is a point that I will return to in later posts: always seek out the toughest criticism of your work as early as possible.
Create your support network
You will need to approach a lot of people for your application, imposing on their time and goodwill: two people for your advisory council, five letters of reference, and I suggest opting for a second mentor for your project. This, for me, is the most important reason to take the time to prepare a carefully considered application. Your future success will depend heavily on your network: surround yourself with good people, and do not take them for granted.
Your co-mentor and your advisory committee members will need to write strong letters of support, but they also commit to advising you through the award. Ideally, all of these individuals will teach you a technique or research area that is essential for your future independence, justifying your request for additional training. As your Specific Aims take shape, think about your proposed research, the weaknesses you have, and how the co-mentor and the advisory committee can complement these weaknesses. You have up to six pages of space for letters from co-mentors and advisory personnel.
You need five letters of reference, which are distinct from the letters of support from your mentors and advisory committee. Be respectful and give your referees at least three weeks to write the reference, but don’t be afraid to remind them of the deadline as it approaches.
I found a co-mentor who is a well-known systems biologist, with expertise in mass spectrometry, quantitative genetics, and integration of large data sets. This was essential to my application -- recall that the first panel identified this as a big weakness, so I needed a strong mentor and a strong training plan for this area. I also recognized that knowledge of molting, and of culturing of the parasite, would be critical for success of the research. To pre-empt criticisms of my own lack of experience on these fronts, I reached out to one expert in C. elegans molting, and another in parasitology, to serve on the advisory personnel. I found my co-mentor at the end of July, and reached out to the advisory personnel in August 2012. I committed to meeting with these advisors along with my co-mentors once a year as part of my training plan. In both submissions the panel remarked on the excellent team I assembled to support my work.
I began requesting letters of reference around six weeks before the deadline, in September of 2012. In some cases you may be asked to draft a letter, which your reference will then modify, should they wish. The letters are confidential; you can see when they are submitted on the eRA Commons website, but unlike the Letters of Support from your co-mentors and advisors, you will not be able to read their contents. I approached my PhD advisor, two lab heads with whom I had collaborated and frequently interacted with at UCSF, and two external collaborators at Texas A&M and CIML in France; I deliberately approached people at a number of institutions in an effort to highlight the reach of my existing professional network. I also tried for at least three references from well-known, well-regarded scientists, and most knew me from my postdoctoral work. One of my references was pre-tenure, which could have been a weakness, but we had collaborated extensively, so the letter was strong and detailed - an obvious plus.
Target your research for a specific NIH institute and contact your program officers
Think hard about the NIH institute to which you are applying. Your science will obviously guide your choice of institute, but the institute you choose should also have a role in guiding and focusing your scientific choices. For example, if you are working on a basic biological mechanism, do you link it to cancer (NCI), kidney disease (NIDDK), focus on the basic biology (NIGMS), or tailor it to one of the other 20 institutes? Look up the funding rates for the various Institutes at the NIH. As you can see in Table 1, there are very different K99 success rates and numbers of grants awarded depending on institution, and each institute will have a specific mandate. Consider the types problems that the institute wants to address, what methodologies they favor, and where they typically send their grant money. Your overall project might not change, but the approach you take to the problem could be adjusted.
Each NIH institute has a program officer; before you write anything beyond the specific aims, contact your program officer. These are extremely busy, dedicated people who want to help you. Send them your specific aims and your biosketch, and they will provide feedback about how the grant fits with the institute mandate and can give suggestions about your chances, potential issues, etc. Contacting the program officer early can save you a big headache down the road. Don’t waste their time and be sure to send something that’s reasonably thoughtful and articulate.
Early on I was faced with a dichotomy: push the research towards the host-parasite interaction, which would be more suitable for NIAID; or focus more on the basic biology of molting gene regulatory networks, which is a better fit for NIGMS; I contacted the program officers at both. The NIAID program officer indicated that the institute only awards a handful of K99s (6 in 2013), so I pushed my work towards basic gene regulation (NIGMS). The NIGMS program officer was kind enough to read my Specific Aims, and pointed out that the reader needed to get pretty far down before finding the developmental and genetic aspects of my proposal. I retained the translational applications, but I re-focused the aims and the proposal to center on basic developmental biology to meet the goals of the institute.
Contact your institutional grants officer and look at other successful K99 grants
You’ve chosen an institute, have some aims, and a deadline to shoot for. Now contact your institutional or departmental grants officer. I was lucky to have an amazing grants analyst. She provided me with a series of deadlines to meet, templates for the administrative material, helped with the budget, made sure that everything was up to code. Your grants officer can also put you in touch with other K99 awardees. My completed application was 89 pages, much of that is administrative material. Give yourself time to do this properly. Seemingly trivial errors could jeopardize your application: even incorrectly measured margins lead to rejection.
I began working with my grants officer about 6 weeks before the deadline. This was sufficient, but I would have liked more time. I looked at four highly scoring (and successful) K99 grants and all had different approaches: some applications have a lot of preliminary data (one had the foundations of a future Cell paper); others were light on preliminary data. I created my own hybrid style. My application was a major departure from my postdoctoral work, so my submission was very light on preliminary data and therefore was more heavily influenced by similar grants. You don’t need to delay submission to generate preliminary data (see the section on the importance of submitting twice). My grant made it to panel review the first time and was awarded upon resubmission with a few proof-of-principle experiments. Panels can and will recognize thoroughly considered, well-planned research proposals.
Writing, resubmission, and some closing thoughts
Give yourself at least six weeks to write the grant, and discuss it often: with your mentors, advisors, peers (and your ever-suffering spouse). I used Evernote to create a K99 checklist, which was invaluable for keeping things on track. I liked to tackle the more challenging parts of the application when I was fresh (Research Plan, the Specific Aims, and the career training documents) and reserve the more administrative things for when tired (budget, equipment, facilities, etc.). After submitting your grant you have an opportunity to update your publication record -- this deadline is one month before the panel meets and around three months after you submit the grant. In some places this is called a “just-in-time” deadline, and in others a “post-submission update” deadline.
For me, the revision period was intense, as was the post-submission update period. As mentioned, I hadn’t published for the first grant cycle, but I had a lot in preparation. During the intervening six months I worked relentlessly, knowing that getting a paper in press would come down to a matter of days. Friends kept telling me that I needed a weekend off. For me the key paper was under consideration at PLOS Genetics, where substantial time was spent finding an editor and reviewers (initial submission) and making minor technical corrections of the manuscript (the revision). The paper was only in revision at the time of submission (July). We resubmitted the paper on September 19th, so there was still a chance of acceptance by the “just-in-time”. The US government closed (that’s right, closed -- full shut-down) October 1, which meant I couldn’t talk to my program officer, because my program officer wasn’t allowed to talk to me. The paper was accepted on October 16th. The government went back to work on October 17th, and I was finally able to get in touch with the program officer. The “just-in-time” deadline was October 18th. I was able to quickly draft a post-submission update letter and get the UCSF contracts and grants officer to sign and send it to the NIH. The notice of award came at the end of April, nearly two years after I began developing the initial grant idea. It’s obviously up to you whether to put yourself through such an ordeal -- I can only tell you about my experience. For me it was worth it.
It’s also worth noting that contrary to popular belief, the NIH isn’t obsessed with work on first-world diseases, directly translational systems, and glamor publications (ie. Cell/Science/Nature). It is possible to get a K99 by publishing in open access journals, while focusing on basic biology, for the purposes of tackling a neglected tropical disease. And remember to thank your program officer.
University of Washington’s Guidelines for Writing Effective Specific Aims
Great blog on general K99 advice (all we know now is that his name is Jake)
How to write a K99 in three weeks (although don't try to do this in three weeks)
10 things that I wish I knew before I wrote my K99 by Tracy Vargo-Gorgola
UCSF Office of Career and Professional Development (OCPD) resources
UCSF OCPD "Preparing a K99" slides (from Bill Balke's outstanding seminar on K99 grant preparation)
Thanks to my spouse -- she says she was suffering but she really was happy to help.