Hi, at this level of detail it is very hard to answer your question, because as you can imagine there are very many tools and whatever is the 'best' workflow depends on many factors. E.g.: as which stage in the genome assembly are you? Is this a new strain, do you want to do a comparative approach? Given that there are many sequenced e.coli strains, an approach that capitalizes on the presence of closely related strains is good. Please provide full details about your goals and precedures, then we can help.

This question has been asked before? http://biostar.stackexchange.com/questions/13238/bacterial-annotation-pipeline/13250#13250

Yes, just put your contigs into RAST (or DIYA) and you might be just fine 90% of the time. Sometimes, when people don't have strong experience with or opinion about a required analysis, then some standard method might as well do, and then explore more options going from there.

You can also submit your sequences to the IGS annotation pipeline (basically the TIGR/JCVI pipeline.

JGI let's you submit to IMG as well, but they are slower. With those two you basically email them and send them your sequence. So you get more support, but RAST is good for a quick annotation.

On another note - I think any pipeline will probably be pretty good for E.coli since it's so well characterized.

Links:

http://ae.igs.umaryland.edu/cgi/index.cgi

http://img.jgi.doe.gov/

RAST, PGAAP and TIGR etc all do a good job but if you want something fast (so you can check, explore) then I have written Prokka http://www.vicbioinformatics.com/software.prokka.shtml