There are two things that affect the quality of an audio file, the file encoding formats, which is what this post intends to describe, and then the actual audio recording itself. It is the latter that PN works on, so the processing that PN does is entirely irrelevant of file type. However, an understanding of the available file formats will help you choose which file format is best for your collection.
There are two main types of audio files, lossless and lossy.
Lossy file formats (MP3, AAC) use algorithms to make the audio files smaller, without a noticeable loss in quality (or so they claim). The simplest way to achieve this is to remove audio frequencies that are beyond the range of human hearing, which is around 20 - 20,000 hz. In theory we can remove all of the frequencies that fall outside of this range without noticing any difference. However, especially on high quality sound systems it is claimed that the loss of lower frequencies can be noticed. Lossy file formats will have a bit rate that represents the quality. 320kbps (kilobytes per second) MP3 files are very high quality, and are debatably comparable to lossless files (although I'm sure there will be posts to the contrary). 128kbps is the lowest that you will usually see music at, although things like audio books that only have a single person talking and little or no music can go much lower, like 64 or even 32kbps. I have heard of some advanced audio encoders which allow fine tuning of the compression that is performed on MP3 files, so that you can choose to have no frequencies removed at all, which would be much closer to the lossless compressed files described later.
Lossless file formats come in two flavors, uncompressed and compressed. Uncompressed formats (AIFF, WAV) are as close to an exact copy of the audio that digital can get, but we pay a price for this level of quality in that they require a lot of storage space. These files also come in varying degrees of quality, 16 bit/44.1kHz files are the quality that the music industry adopted long ago for Cds. There are also 24 and 32 bit versions, and sample rates that go all the way up to 192kHz but these ultra high quality files are usually only used in music studios for sample content or other special use cases in the music production process. Again, I'm sure that there will be posts to the contrary but there is no reason for the average DJ to use files that are higher than the industry standard 16/44.1 format. WAV files use a poorly supported format for metadata, so should be avoided in favor of AIFF whenever they are available.
Compressed lossless formats (ALAC, FLAC) use compression algorithms to make the file size smaller without any actual loss in quality. Lossless compressed files offer an excellent choice for good quality and moderate file size, but generally speaking are not as widely supported. ALAC files share the same file extension as AAC (.m4a) and are probably a little bit more widely supported than FLAC files but I don't have any empirical data to support that. If you're wanting to use one of these formats you should make sure that whatever software and devices that you want to play them on supports the format that you choose.
So which output is best in PN? It depends mostly on personal preference, but if you are exporting to MP3 you should be aware that there's a tiny bit of data loss every time that you encode a file to MP3. You won't be able to notice it until a file has been encoded multiple times though, and one MP3 to MP3 conversion won't be noticeable. My opinion is that if you are worried about the data loss from a second MP3 encoding you shouldn't be using lossy files at all. Unless you have massive amounts of hard drive space it's not worth turning a lossy file into lossless, unless either you want to perform further editing on the file in the future, or if you don't want to keep the originals (we do recommend backing up the original files) it's better to have a lossless version of the file as opposed to an MP3 copy as the only version of the song in your collection. Creating a lossless version from a lossy source file does not improve the quality of the underlying audio, it just creates an exact copy of the lossy audio, without any further data loss. Just remember that best practice is not to make copies of MP3 files.