Dear friends and colleague,
I just want to work on cancer especially (triple negative breast cancer) early diagnosis and detection. So that I just want to analyize cancer data for finding markers. But I dont know how to do that. Initially I just want to indicate that I am newly graduated student holding BSc. And now I have plenty free time and I want to learn data analysis. I just tried different things like R, Python, Knime, Orange and Weka. But I didnt find a outline or pipline what I needed to follow during analiysis.
Please help me or send me some articles how to do that.
I am strongly interested in DNA methlation changes as well. Maybe even my question is irrelivant.
I am strong interented in computational biology and bioinformatics but I dont know where to start and proceed.
Thanks in advance.
Hi Suleyman,
One thing you could do is to start with a relatively simple problem first (I would recommend this approach). To get started with some data analysis, you could use some simple data from (http://archive.ics.uci.edu/ml/datasets/Breast+Cancer+Wisconsin+%28Original%29). This is a problem that I typically give to my students starting on machine learning and data anlysis in comp biology. With this dataset you could try to predict breast cancer diagnosis/prognosis with the python scikit library (http://scikit-learn.org/stable/). For example, you could write a code that uses a Random Forest to analyze the dataset and make the predictions. This would be a good exercise for attempting more complicated and larger cancer/bioinformatics problems. The essential ideas, however, will remain the same.
In terms of Epigenetic data, I would suggest first taking a look at the collection in Nature. http://www.nature.com/collections/vbqgtr. This is the epigenomic roadmap collection and you will find papers related to detecting "promoters and enhancers". Also, the epigenome project has made lots of data available that can be freely downloaded (although they are huge files). An open problem is to accurately identify promoters and enhancers that correlate to different disease phenotype.
Good luck.
Hi Suleyman!
I would recommend you to take a look at the Data Science Specialization in Coursera. You can take it without certificate for free. Reading their books alongside the classes is a good idea. It's a good introduction to a Data Science "mindset", and fundamental steps. The more advanced statistics (regressions, machine learning) are not so clear, but if you have some prior knowledge or take other online courses on the topics, you should be OK.
The same professors have courses in biostatistics, bioinformatics and genomic. If you have time you can take the first courses fast, and then select which of the others are priority.
Cheers!
Download my several papers after last October. I can find the structure of microarray datasets compleately and find cancer gene subspaces.
I published two books.
#New Theory of Discriminant Analysis after R.Fisher (Springer) # introduces how to find several disjoint unions of cancer genes (Small Matryoshka).
#From Cancer Gene Analysis to Cancer Gene Diagnosis (Amazon Kindle version, $9.99, Unlimited user free)# explain how to analyze SMs and find cancer gene diagnosis.
- Badges
- Science topic