Re: Cluster analysis for beginners
- From: "illywhacker" <illywacker@xxxxxxxxx>
- Date: 29 Mar 2007 02:41:36 -0700
On Mar 29, 11:17 am, Sidney <milan_y...@xxxxxx> wrote:
Dear all,
I ran into a problem that I can't solve with my basic statistics high school knowledge - apologies if this is too trivial to some of you.
Assume you have 5000 proteins that are ordered by their molecular weight from 1000 Daltons to 100000 Daltons (the numbers don't matter). If you now find that a certain motif (e.g. a specific phosphorylation motif) which is only found within a certain molecular weight range, e.g. only between 77000-81000 Daltons, how do you determine if this 'clustering' is significant? At this point I have no idea what to do and where to start at. Your input is very much appreciated.
-----------------------
The first point is to be very wary of the replies that you will get to
your question. There is no such thing as a generic statistical test of
significance, and yet that is what people will suggest to you. Almost
all replies will involve strong assumptions, equivalent to underlying
models that may or may not be acceptable. Unfortunately, these
assumptions will not always be made clear to you, often because your
respondent will be unaware of their presence. In addition, classical
statistical methods of testing are deeply flawed in general.
The first question is: what do you mean by "significant"? What do you
hope to find out/prove? Without knowing this, your question is not
clear.
illywhacker;
.
- References:
- Cluster analysis for beginners
- From: Sidney
- Cluster analysis for beginners
- Prev by Date: Cluster analysis for beginners
- Next by Date: Re: A letter to a statistician. Emailed and airmailed and no reply
- Previous by thread: Cluster analysis for beginners
- Next by thread: Re: Cluster analysis for beginners
- Index(es):
Relevant Pages
|
|
