I have 10,000 fixed length strings of DNA sequences. An example would be
ATTGGGT M GCGGCTG. Now the character marked M is a position of interest to me (say something that causes diseases). Similarly, I have data for non-disease sequences also where the position of interest doesn't cause a disease.
How can I say that the neighbouring sequences like A,T,T,G,G,G,T to the left or G,C,G,G,C,T,G to the right have an influence on M being a disease-causing/non-disease causing position in the sequence.
How can I conclude something like T in the immediate left of M has more influence than A in the seventh position to the left of M.
I thought of using regression to predict M using the neighborhood sequences but I am not sure whether that would work. Can I use some clever encoding techniques to prove the above hypothesis? Any help is greatly appreciated.