most basically, an NPN is used to supply a path to ground (called a low-side switch) and is turned ON when the signal to the base is HIGH (specifically, higher than the voltage at the source terminal). A PNP is used to supply a positive voltage (called a high-side switch), being ON when the base is LOW or grounded (specifically when it is lower than the source terminal).
In most cases, the best PNP transistors are less efficient than the best NPN transistors, so many people use NPN's as high-side switches, using a voltage doubler or charge pump to bring the gate voltage above the source voltage so the NPN will turn on. This however, is more complicated and requires more parts than a simple PNP high-side switch.
Huh? PNP/NPN are pretty symetrical, there isn't much difference between them, and they have base emitter junctions, not gates. You are confusing them with MOSFETs, where N-channels are more efficient.
In some limited circumstances BJTs and MOSFETs can be swapped, but the theories of operation are totally different. Most MOSFETs require at least 10V to turn on fully, though there are versions that can use less. They are voltage devices.
A BJT (the type of transistor the OP is referring to) uses current. A general rule of thumb is 1/10 the base current is required to turn the transistor on fully, although there are lots of exceptions to this too.
To the OP, if you give us a specific example how you want to use them we can be of more help.