I sympathize, and I wish I had more specific suggestions.
>problem. My sponsor is a theoretician and has little lab experience.
>Also she has notoriously poor social skills. She has been very generous
>with money, but otherwise a disaster -- mostly ignoring me, sometimes
>giving bad advice, and the day before a grant proposal deadline asking for
>all kinds of results that she never cared about before. She has even
>ignored questions posed point-blank by e-mail, which is, in fact, her
>preferred form of communication.
This is quite frustrating, I can imagine. Perhaps you have to ask the
questions more than once, or first warm her up by email and then catch
her later, verbally, "oh, did you get a chance to think about what I
asked you yesterday about PCR?"
>>What should I do? The people here have been generally helpful, but no one
>can give the project the kind of oversight that it needs.
I'm not sure how your sponsor would react to this, but what about seeking
out an "expert," even at another institution and developing some kind of
relationship with him/her that you could ask questions.
My project got pretty far out of my and my advisor's area of expertise,
from neural development into glycobiology. To learn more glycobiology,
I subscribed to bionet.glycosci, and went to a few glycobiology meetings
(one on the net and a Gordon Conference). I've met a few glycobiologists
this way and will ask them questions over email or the phone.
I have also just found people who could be helpful by reading the
literature, looking them up, and sending them a letter. Sometimes the
letter drops into the void and you never hear from the person again,
but sometimes you can even go so far as to have a collaboration and/or
publication with a person you approach. In my case, I've found it's
helpful that I have some neuroscience knowledge to offer in exchange
for the glycoscience knowledge. Sometimes
I do feel a little awkward, but people have been generally generous and
responsive. I do have to swallow my pride and just go ahead and ask
stupid questions. I sometimes preface these with something like "I hope
you can help a struggling neurobiologist who doesn't have much background
in this area," or "I'm sorry to sound so naive," but I also realize you
don't want to say too much of that.