Help with a molecular biology project

#1
Hello guys
I am doing my phd research in molecular biology. I am totally illiterate in statistics that is why i am facing some real problems in my proposal.
coming to the point, below i describe my project in simple words. i need you guys help and comments and suggestions for modifications
I am working on correlation of a certain abnormality in DNA of patient with hepatitis c related cancer. my hypothesis is that these changes start occuring in dna at very early stages, long before the cancer actually develops (HCV infects chronically and people remain infected for decades before they actually develop cancer. the disease progresses step wise from simple chronic infection to cancer). wat i want to do is to obtain the samples of patients at various pre-cancerous and cancerous stages and compare them for the changes in their dna (these changes have been reported to be involved in cancer development). this study will be a little biased as we will be taking only HCV infected patients and that too from a certain locality visiting a certain hospital.
now my question is:
wat would you call this study .... cross sectional, case control.... or wat else? how would we define the study design and sampling methodology?
secondly, wat statistic tests/analysis need to be done to analyze this data, one thing that i have seen in published research of this kind is that they either use fisher's exact test or in some simpler studies they used mean SD. i simply dont understand statistics so i can get help from these papers.
I need to submit my proposal urgently otherwise i wud have educated myself before venturing into this area

Looking forward to some guiding comments. I would welcome if someone can point out flaws in my study design. and lastly it would be a real help if someone can provide me ready made answer

Thanks
 

fed1

TS Contributor
#2
Just glancing through what you have written here I want to urge you to follow subjects prospectively.

If you show that Subjects with cancer have a greater chance of having the mutation, little is accomplished, because we do not know when the mutation developed. This might just be ignorance on my part of how this mutation works.

Following subjects prospectively means:

1) Find dudes with and with out the mutation.

2) wait.

3) Tabulate the proprotions getting and not getting cancer. You should be looking at a 2 by 2 table.

4) Compute relative risk and confidence intervals.

Gene trait association can get hairy when the muation is at multiple loci and we do not know phase allignment, but I am feeling like this is a simple mutation.

Hope that helps and keep us posted on your progress. :)
 
#3
well i must tell u that im not looking for a mutation. i am looking for gene promoter hypermethylation. this is very much in contrast to the mutation thing. mutation is kind of yes-no condition. but these methylation changes are sort of quantitative thing, progressively increasing as the disease progresses.
secondly, considering the thing that i have to complete my phd in a specific time i can not propose a 20 yrs long study. that is why wat i thought while planing this project was that we take a snap shot of the "things". patients from all the different stages of disease be compared with each other to have an idea about possible methylation alterations.

I hope you get my point