1. Yes (but applied physics, stats, CS or engineering is more valuable) 2. Yes and no.. if you are not quantitative then yes. 3. Yep quite a few that are more bio than quantitative. 4. This is individual really. Many places have Directors of Bioinformatics. In my experience these people are usually statistically trained or algorithmically trained people. Good at math.

So pick the angle you want. Bioinformatics can be statistics, algorithms, big data, genetics etc...

From a CS perspective, almost everything is a graph and then you have sequences as well. From a stats perspective everything is a noisy data from some model and from a engineering point of view it is signal processing.

Otherwise you are just creating perl/python scripts to munge data into other people's pipelines. That you want to avoid. I think a PhD is really the only way to get around it.

Look at the big tools in the industry and the background of the people who created them.

Lastly you still need a biologist or someone with a bio background to design the experiments and interpret them!!