home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!mcsun!uknet!comlab.ox.ac.uk!oxuniv!oxpath!rhubner
- From: rhubner@vax.path.ox.ac.uk
- Newsgroups: sci.math.stat
- Subject: test for random association
- Message-ID: <1992Sep14.231604.1@vax.path.ox.ac.uk>
- Date: 14 Sep 92 22:16:04 GMT
- Organization: Oxford University Molecular Biology Data Centre
- Lines: 67
- Nntp-Posting-Host: oxpa01
- Nntp-Posting-User: rhubner
-
- Hi everybody,
-
- I'm writing up a "D.Phil." in Zoology here at Oxford and I'm stuck with the
- following stat problem about chromosomes that join together.
-
- The background:
- 1) mice have usually 2n=40 (which means 20 pairs of chromosomes)
- -> these chromosomes look [schematically] like this:
-
- *
- a / \
-
- 2) some populations 'fuse' the chromosomes of two pairs
- -> looks like this and is written as Rb(a.b)
-
- a \ / Rb = Robertsonian fusion or centric translocation
- % % = **, but now they segregate as one chromosome
- b / \
-
- 3) a maximum of 9 fusions (2n=22) is known from some populations
- (leaving one pair and the sex chromosomes unfused)
-
- ex: Rb(a.b), (c.d), (e.f), (g.h), (i.j), (k.l), (m.n), (o.p), (q.r)
- with s,t, X,Y unfused
-
- 4) the same chromosome combination can be found in mice with 1 fusion (2n=38)
- or mice with more fusions (2n=36-22)
-
- ex: 2n=34; Rb(a.b), (c.d), (e.f)
- 2n=26; Rb(a.s), (c.t), (e.f), (g.h), (i.j), (k.l), (m.n)
- ^^^^^
- 5) fusions are not independent, because they are sampled without replacement
- from the initial pool (N) of 20 different chromosomes; this reduces the
- number of free partners after each subsequent fusion
- -> the number of different combinations [written C(N,k)] for a chromosome
- set with a given number of fusions (k) is thus
-
- N!
- C(N,k) = ------------------ k=1,2,...9
- 2^k * k! * (N-2k)!
-
- The question: How to test for random association given that the probability of
- a fusion in a certain context should be quite different? (Some people think
- that I should just take 1/C(N,2) as the expected value, but I believe they are
- wrong). Furthermore, because I may not be absolutely sure about the sequence of
- fusions in a particular population, is there any other way to test for random
- association here? Does somebody know about analogue systems (lottery?) and
- point out where the solution may have been published? Or is this a totally
- uncommon and 'desperate' situation?
-
- Thanks for any hint!
- Roland [rhubner@ox.path.ac.uk]
-
-
- .... ....
- ' )._._._( `^^^^^^^^^^^^^^^`.
- ( ( ) ) .\
- \. \ 0 0 / ./ . )\\
- `._/\ /\_.' ) ( / \\
- ==&__&== `_| / _._._.( _/ \\
- "" /__| /_._./ `*****
-
- ***********************************************
- * Roland Huebner, Dept. Zoology, Univ. Oxford *
- * South Parks Road, OX1 3PS, Oxford, England *
- ***********************************************
-
-