Life, of course, being basically a piece of source code that has instructions for reproducing itself. One can apply other standards to life and get other results, but this is basically the definition I'm using for now. The funny thing is that one begins to notice that some religions are essentially 'life' - they are pieces of source code with instructions to reproduce themselves embedded. I'll talk more about this some other time, because I think it's important for you religious types out there to understand that you may have been duped, but for now, we were talking about life.
Anyway, the discovery of life from extraterrestrial origin of course has lead many people to conclude that maybe life didn't originate on this planet. While this is probably true and certainly interesting, I don't think it's the biggest question one might want to ask.
One of the first things we might observe about 'life' - that self-reproducing and presumably self-modifying source - is that the odds of it 'just happening' are finite but very, very low. That is to say, of all the random combinations that the various chemicals that make up the instruction set could land in, the odds of them landing in the right pattern to make reproducing and improving source code 1 in a number so large it probably won't fit on this screen. But they _are_ finite. Which means that you still have the choice of postulating a creator or not postulating a creator, as you choose.
If you do postulate a creator, of course, then you have to ask where that creator came from. And no, the answer 'well, He/She has always been there' won't wash.
One of the recent advancements that may offer quite a bit of insight some day is the human genome project. We've finally mapped out the entire source code of one human - but don't expect them to start telling you which lines of source make your fingernails crunchy tomorrow. You see, there's a little problem here. It's not insurmountable, but it's kind of big.
The problem is thus: think of the human genome as a program. Instead of running on a computer, it runs on a human cell - and instead of being stored in magnetic north and south orientations, it's stored in four acids - but none the less, our DNA is a program. The problems is that like any program, it's in machine code.
That doesn't sound like that big a deal - after all, with enough practice, a really good hacker can read machine code. But it actually is a rather big problem - in that not only is the DNA in machine code, we don't know the _instruction set_. We don't even know how big instructions are, or what arguments they take.
What we need is to brute-force crack the instruction and data sets of the human genome and write a reverse-compiler. It need not go to a high level language - going to something like assembly would probably be good enough. The reason we need to do this is that we could probably learn a _lot_ by reading the human source code. (One religious friend of mine calls it 'the fingerprints of god'. I'm not sure if I agree, but it's certainly going to be interesting).
The biggest question, of course, is what are all the sections of source that we _don't_ use? Do they all even have the same instruction set? It's possible that half the human DNA - or even considerably more - is in older instruction sets for older designs of cells.
This of course increases the difficulty of cracking the instruction and data sets - one of the first things that might be necessary is to figure out what section of the source is currently 'live' and what sections aren't. This might be sort of tricky, but I'm sure the bright boys that do the research could figure out some way to do this, if they haven't already
Another interesting thing about the DNA is that it must, by definition, contain instructions for building the interpreter - so once we figure out the instruction and data sets, we'll be able to read the source for a human cell - this may also prove to be enlightening.
Certainly all this is worth doing. Certainly it may take many, many years to figure out the instruction set and document it - it doesn't help that each 'bit' can have four states instead of the more usual two - nor does it help that the area we're trying to map is so big. But, once we've done all this, THEN we can start debating what similarities the bacteria from outer space and life on earth have - because once we've done this.. twice, once for earth life and once for the ET life - then we'll be able to search for similar instructions and visually inspect the code that generates each type of cell. Until then, we're just guessing.
And who knows? The human genome might contain a exact copy of ET's cold germ's source embedded somewhere in it. After all, the human source code is _big_ - it may contain many library routines that never get used any more - because hey, the space is free. Or it may not. But either way, it's too early to conclusively argue over where we came from.
It does seem likely to me that the odds of that self-reproducing source we call life occurring two different places are pretty low. And, it is certainly possible that the bacteria that landed here many millions of years ago 'seeded' the planet, thusly starting the life we know today. After all, once you get the self-reproducing self-mutating code executing, time will result in better and better life forms. It's the starting point that's the hard part.
But I also will point out that we don't really have enough information to know. Someone who's really bored might figure out what the absolute shortest set of instructions that would result in self-reproducing self-mutating life would be, and then what the odds of the acids lining up to produce that set of instructions are. I imagine the odds are lower than people tend to guess.