Analysis Of The Cancer Genome Atlas

Question

[Pappu]

I am a newbie experienced in bioinformatics, linux and python programming interested in cancer genomics. I don't have any funding at the moment to sequence cancer genomes from patient samples and publish results. So I am wondering if it is a good idea to download cancer genomes from https://tcga-data.nci.nih.gov/tcga/ or some other source (let me know if you are aware of any), perform some preliminary analysis using different variant calling algorithms, publish some short papers and then apply for funding. I have access to several clusters and supercomputers for processing data. Thank you.

[Giovanni M Dall'Olio]

Of course it is a good idea to use data from public resources, but you will have to plan your project very well before starting.

Take into account that for the data in the Cancer Genome Atlas, other people have probably already done a variant calling analysis. The best thing to do is to contact each institute directly, and ask them if you can contribute in any way, and how.

[Pappu]

Thanks for your suggestion. I will write to them. I just gave an example of variant calling which is very tricky. Given the massive amount of data, there will be plenty of possible analysis. I wonder if they will be interested in me. I need to show some preliminary results in order to get some funding to run own sequencing projects.

[Chris Miller]

The only thing you need to contact TCGA for is permission to access the restricted data (which is access-controlled for patient privacy reasons). They'll want to see that you have some kind of plan to do cancer genomics research with it, but it can be loosely defined. The point of the project is to produce data that will be available to the scientific community.

[Pappu]

I think everyone's plan is more or less similar - to identify novel mutations, genes or biomarkers for that particular type of cancer.

[Chris Miller]

You'd be surprised at the variety. There are lots of people working on things like algorithmic improvements or pan-cancer analyses, for example.

[Pappu]

Yes I was reading about the RNA-seq Genome Annotation Assessment Project.

Answer 1

[Sean Davis]

I would agree with Giovanni that your best bet is to offer your services to a group actually working in cancer genomics and start from there.

I have found that a good way to market your services is to share your knowledge via classes for local researchers involved in such work. A cancer center, department, or lab/division will typically jump at the chance at free instruction and education.

[Sean Davis]

This isn't meant to discourage you from getting access to TCGA data; you can do that regardless of having cancer genomics collaborators, but you have skills and often cancer genomics groups have interesting problems.

[Pappu]

Thanks for your inspiration. People who are doing NGS already have some bioinformaticians in their lab or collaborations. Or they are interested in collaborating with a big guys- not me. I don't have any publication in this area/tenure track position to attract collaborations. I told some people working on cancer to sequence exomes of patient samples and do RNA-seq. But they are scared to do so due to lack of data analysis skills and happy with western blotting.

Since last two years, I have been interested in NGS data analysis. So I read many papers in the field, followed several tutorials, did analysis of the 1000 Genomes and complete genomics data etc. Also I gave several postdoc interviews in different western european nations, and had technical discussion with experts in the area who are publishing good papers. So I understood pretty well what they are doing. I actually had a postdoc positon in a leading bioinformatics group for such NGS data analysis. I rejected that to stay in an experimental group to generate my own data.

So the bottom line is, what I can offer is biologically meaningful data analysis or try to find something which has biological significance.