PREVIOUS ADVICE 
I'm a biochemist studying structural biology with a new research faculty position at a small-to-midsize university, the result of a spousal hire. I have my own small, unfurnished lab space--but no research money. My work straddles the NIH and NSF worlds, so funding from either agency is a possibility.
I need seed/start-up money to get the preliminary results that NSF and NIH are going to want. Some funding agencies claim that preliminary results aren't required for new investigators, but I've been told this isn't true.
Any thoughts on small start-up grants that don't require a tenure-track appointment? I hold a 3-year renewable appointment.
Your situation is difficult, but full of potential. On the one hand, the lack of startup funds is a serious problem. Your lack of postdoc experience--and the publications that normally come from a postdoc--will not make getting money for research easier. On the other hand, in giving you lab space and 3 years' salary, your institution has demonstrated a real commitment to you. It's not as good as a tenure-track position, but from a funding standpoint, it's better than what many young researchers have to deal with.
Still, one of your biggest challenges will be to convince potential funders of your institution's support. You need a letter from your administration stating the strong commitment of your department and university to your career development, and assuring the funding agencies of your guarantees of lab space, institutional salary support, and access to all core facilities, for a renewable period of 3 years. A statement confirming that if things go well (for you and your institution) your position, conceivably, could be converted to tenure-track, would be very useful. But even without such a statement, the fact that the university has committed to 3 years of salary support will be a major plus in the eyes of funding agencies. You are far better off than the typical post-postdoc .
Speaking of institutional support: start your funding search within the institution. Many colleges, including small ones, have internal pots of money to help support research projects and student research experiences. Support from such a fund might be just what you need to pay for minimal travel (to support collaborations) and basic supplies and reagents.
Next, what you need is to find at least one collaborator/mentor, preferably nearby, who can advise you and share not only expertise but also facilities. With arrangements like this, researchers at smaller institutions--people who have no more startup money than you do--often manage to do great science. And while they have the advantage of a tenure-track job, you have the advantage of minimal teaching obligations.
Utilize these collaborations to get the preliminary data you need. With strong letters of support from your collaborators and institution, apply for modest research grants from NSF, NIH, and other suitable public and private funding bodies. NIH's review process doesn't include close scrutiny of budgets, so you can use that money to buy equipment, though smaller grants won't pay for big-ticket items. In time, you can utilize infrastructure-improvement grants to buy the more expensive stuff, as well as educational-infrastructure grants to buy equipment that will serve both education and research.
Since you are at a small university--which, presumably, doesn't get a lot of support from NIH, you probably qualify for the R-15 program, the Academic Research Enhancement Awards. If you write a good proposal, the odds of getting an R-15 are excellent.
Propose to study a problem of real scientific interest, demonstrate a thorough awareness of the relevant literature, and prepare a work plan with a promising scientific approach. Plan the work so that subsequent aims don't depend on the successful completion of some early high-risk step. Get other scientists--including people who serve on grant-review panels--to read your proposal and suggest revisions. Put it through several stages of substantive revision and editing. Make it precise and snappy, emphasize the big picture, and provide just enough detail to make it compelling and convincing and to demonstrate intimate knowledge of the subject. Crucial, difficult steps should be informed by evidence from the literature outlining approaches in similar systems that already have been shown to work.
If it isn't funded the first time, revise carefully and resubmit.
Best of Luck,
I am about to receive and activate a NRSA fellowship. When I wrote the fellowship, I expected to spend all my time working on the project. So I put 100% effort in the application.
But now, I am thinking of submitting an RO3 with a colleague in the lab on a short, defined project that we can both work on concurrently on the side of our main projects.
Am I allowed to do this? Can I change the 100% effort for my NRSA to 80% and spend 20% on the RO3 project?
The short answer: No. You cannot simultaneously hold an NRSA and a research grant, an R03 or any other government research grant.
Employee postdocs are still considered to be trainees; if we didn't consider them trainees, how could we justify paying them so poorly and treating them so badly? But from a legal standpoint--including considerations of taxation, etc.--their trainee status is ambiguous. Like so many things, it all comes back to tax policy. Employee postdocs are compensated for their work. They pay income taxes, including social security and unemployment tax.
Postdoctoral fellows, however--and this includes NRSA recipients--are unambiguously considered trainees, not employees. Indeed, NIH forbids universities from treating NRSA recipients like employees (which is why you can't get health insurance via the employee health plan). It may come as a surprise to you, but your stipend is considered a grant, not a wage; it is not payment for services rendered. And because it is not a wage, it cannot be subdivided, as wages can and often are. You cannot attribute 80% of your time--and take that portion of your salary--from your NRSA stipend, because your NRSA stipend is indivisible.
Furthermore, though institutions are allowed to supplement your NRSA stipend (as long as you continue to work at least full-time on your training project), they are not allowed to supplement your stipend using DHHS (Department of Health and Human Services) funds.
If you were an "employee postdoc" and not a fellow, you would have none of these restrictions, because your status (trainee or employee) would be adequately ambiguous in the eyes of the (tax) law.
Nothing prevents you from applying for an R03, along with your colleague. But if you succeed in getting the grant--don't count on it, since success rates are falling like rain--you will have to give up the NRSA. So make sure the grant includes enough money to pay all your salary.