These choices depend primarily on the quality and characteristics of the data that you need to process and will not depend on the organism in question. The reason that these are not filled in already for you is that no one but you can determine what the right approach is.
The first two concepts have direct statistical interpretation. The third is about dealing with missing values, you could remove missing values or replace them with averages or other estimators. But again you would have to think it through and/or experiment with choices.
In general I would not worry too much about choosing the perfect parameters, it has been shown that strong signals are not affected by a particular choice of normalization, whereas trying to detect subtle effects without understanding the strengths and weaknesses of the methods is not likely to be successful anyhow.
In short try experimenting with various approaches to get a sense what the type of results they lead to.
