What flavour of quarks make up quark gluon plasma? What would it most likely cool down into, protons, neutrons, strange matter, neutronium or what?
Quark-gluon plasma is just a phase matter in QCD may enter into.
Just as a solid or a gas may be made out of any kind of atoms or molecules, quark-gluon plasma may be made out of quarks of any flavor. If you make it by heating ordinary matter, it will obviously mostly contain up- and down-quarks. It's just a phase of QCD, and it cools into whatever its constituents can cool into.
There are two conserved charges, baryon number and electric charge. This means that the total baryon number and charge of a quark gluon plasma are determined by how it was made. Everything else is determined by thermal equilibrium (if there is enough time), and kinetics (if there is not enough time to fully equilibrate).
In a heavy ion collision there is not enough time for weak interactions to equilibrate, so net flavor (up minus anti-up, down minus anti-down, strange minus anti-stange, charm minus anti-charm, .. ) is effectively conserved as well. In the quark core of a neutron star (if such a thing exists) weak equilibrium is reached.
In a relativistic heavy ion collision, as a first approximation, you can think of all quarks as heaving been pair-produced by gluons. This means that the QGP has no net flavor or baryon number. Of course, the initial nuclei had net baryon number and a slight excess of neutrons over protons (i.e. down over up). Most of this goes down the beam pipe, but some of it ends up in the plasma, so there is a slight excess of quarks over anti-quarks, and a slight excess of down over up.
If you look at the total number (up plus anti-up, down plus anti-down, ..) rather than the net number (up minus anti-up, ..), then in thermal equilibrium there will be a tiny excess of up over down, and a larger excess of up over strange, because of the mass ordering of the quarks.
Experiments with relativistic nuclear (heavy ion) collisions at CERN-SPS in '90s at BNL-RHIC in first decade of this century and 2nd decade at CERN-LHC have all produced quark-gluon plasma consisting of gluons, up-quark pairs, down-quark pairs and strange-quark pairs. At LHC highest energy also protons can induce formation of QGP fireball, and in nuclear LHC collisions we see lots of charm pairs, the 4th quark flavor, and even bottom flavor (5th flavor).
At the SPS (fixed target) this new phase of matter shows high additional up-quark, down-quark contents (not pairs) relating to a fraction of baryon number originating with the colliding nuclei being stopped in the fireball. Since charges of up and down are not opposite, aside of baryon number (three quarks=one baryon number) there is also a net fireball electrical charge. These effects are smaller at RHIC (collider) which reaches 15 times CM collision energy of SPS and are unobservable at LHC where the energy is up to 25 times higher compared to RHIC.
There is no time in these collisions to convert any flavors of quarks into each other by known weak interaction processes: the QGP fireball explodes too fast, explosively hadronizing i.e. creating matter out of energy of the primordial -soup, comprising many free streaming new particles helping us understand QGP properties. A YouTube channel collects the related clips on creation of matter https://www.youtube.com/channel/UCRYj1UAYtCfXQdLeeiENu8A
The quark model has developed to fit a great number of measurements
In particle physics, the quark model is a classification scheme for hadrons in terms of their valence quarks—the quarks and antiquarks which give rise to the quantum numbers of the hadrons. The quark model underlies "flavor SU(3)", or the Eightfold Way, the successful classification scheme organizing the large number of lighter hadrons
Hadrons are not really "elementary", and can be regarded as bound states of their "valence quarks" and antiquarks, which give rise to the quantum numbers of the hadrons. These quantum numbers are labels identifying the hadrons, and are of two kinds. One set comes from the Poincaré symmetry—JPC, where J, P and C stand for the total angular momentum, P-symmetry, and C-symmetry, respectively.
The remaining are flavour quantum numbers such as the isospin, strangeness, charm, and so on. The strong interactions binding the quarks together are insensitive to these quantum numbers, so variation of them leads to systematic mass and coupling relationships among the hadrons in the same flavor multiplet.
Looking at the eightfold way of calassifying into flavor SU(3) the measured resonances/particles we note that one of the axis in the plot is proportional to the mass of the resonances.
baryon decuplet -------------------------------- meson octet
The strangeness equal zero state has the lowest masses in both. The quark gluon plasma happens at very high energies, proposed in the hypothesis for the evolution from the singularity of the Big Bang as the energy falls due to the expansion of the universe. As the plasma cools there will be all flavors (and antiflavors) according to the available energy to generate the masses , further cooling will disfavor the hadrons with strange quantum numbers, which will follow the normal decay paths, until protons and neutrons can form and the plasma stops existing, at about a microsecond from the Big Bang.
The other way, trying to create a quark gluon plasma is being studied at the LHC at the moment .
So the quark gluon plasma does not have a unique flavor but is a statistical distribution of flavors, of elementary particles and antiparticles , depending on the energy available and the quark model. From conservation of quantum numbers in the creation of flavors in pairs of particle antiparticle , the overall plasma is flavor neutral.
Well it really depends if the things you heated up had mostly nucleons then you can expect a lot of up and down quarks. However because of the Heisenberg's uncertainty principle you can never truly know what particles are inside a proton or neutron. Instead all you know is that there are three valence quarks(the quark composition for proton being up, up, down and quark composition for neutron being down, down ,up quarks). So there is a really low possible you could get a charm quark as long as there is a anticharm quark somewhere. It is okay to have an antidown quark as long as there is an extra down quark on top of what is needed for it to be a proton or neutron. If there is enough energy all these virtual quarks can become real. Then the resulting particles can be anything really. It does not have to end up as protons and neutrons. Instead they can end up as Delta particles. However these particles have a very short lifetime and all decay into protons and neutrons. Even neutrons are unstable on their own and will decay into protons in 15 minutes. There is a also the possibility of pion and antiproton production. Though pions are unstable and antiprotons annihilate with protons into a shower of other particles.