Single sample composing

I think I get what you mean, although technically a single sample is a number, say, something between zero and whatever bit-depth you are using. So I take it you mean a sound file and maybe a single sound within that sound file? Like a blip or flute playing a note or an explosion. Or it could be an entire sound track to an old tv episode of My Mother the Car? Chopping up a sound file that is a sequence of sounds--such as a tv episode or a Schubert symphony recording or a song by Iris DeMent--has been so BTDT [been there done that], either literally in the analog world with a non-magnetic razor blade or digitally with a software blade, so I am further guessing you mean the single sound variety. Although, still, an explosion or an organ chord (including the reverberation tail) is so much more to work with than a blip from an Asteroids game.
 
Okay, given a relatively simple sound file, like a flute note, there's always the "cheat" method. Put the sound into, say, Dim Pro and concoct a flute melody. Or a quartet of digital flutes playing a movement from Scarlatti. Or 2,048 flutes, each one tuned 0.001 hertz higher/lower than the next, playing an all-flute arrangement of Ode to Joy.
 
To get to the "purist" version of your exercise? I tried it once, got to six tracks (in ten minutes or so) using one of the Oscar sample files under BigTone in, um, Rapture's multisamples (I think), found it extremely tedious going and that was that.
 
If I was to try my hand at it again, I would start with a software device that focused more on granular synthesis (like Alchemy, RIP, for example) and less on the filters (like Rapture). Probably better results by moving generated sounds back-and-forth between multiple synths, to take advantage of each synth's strengths? Lots of filter ideas spring to mind (I tried my experiment a couple of years ago, I'd have a much broader set of ideas to try now). Oh heck, there are so many ways to go, now that I try to list.
 
Actually, an approach I might try would be to take one of those apps that takes a sound file and continually morphs and mangles it over and over (for instant background noise--er, I mean, "sound installations"), record a couple minutes of that--THEN chop up the results in a DAW and try piecing things together. Programs I've seen are "replay player" or "crusher-x", don't know if they still exist. I haven't seen ARAFM (automated random audio file manipulation) as the next big thing in Wired or anything. None of the handful of programs I've seen have been very creative, they kind of pick an algorithm (hah, don't we all) and go with it ad infinitum. A really creative program would use: granular, filtering, tempo, pitch, direction, rearranged sequencing, combining, subtracting, temporal adjustments--in short, ALL the many ways we dumb, but clever, human computers would come up with to change the audio output of that one dorky ol' sound.
 
If you decide to go on with your exercise, I don't think you'll run into a bunch of sound-alike competitors. (Although I guess there's probably a forum out there somewhere....) Like I said, it's tedious work if you do it by hand, and it's boring if you assign a machine to do it.

I got one of the Anomaly Alchemy soundset which use the same samples in different synths for similar presets in Kontact, Alchemy and Rapt (or DimPro?).  I did about 3 tracks of the same noiseful sounds using the collection.  No tuned sounds, it is rhythmic noises.  Similar sounds but not the same. The piece will be called 1st Date and it is for a narration about a, well, first date while tripping.  So the music is intentionally inharmonic and spacey.
 
@

I have heard songs using nothing but the 808 kick tuned, filtered and mangled to death to provide rhythm and melody. 

That sounds awesome - I love homemade contact mics, never thought of using them like that though, giving me ideas...
 
Time is an issue, always, but sounds like you have a plan for it and that always makes me want to get to it more quickly and makes things happen more quickly when I get there.