The 21 cm (1420 MHz) line of neutral hydrogen is the result of the electron in a hydrogen atom flipping from positive to negative spin. When that electron drops into the slightly lower state, a very low energy photon in the microwave region of the spectrum is emitted.

This occurs in monatomic hydrogen, not the more familiar hydrogen molecule (H2) we are likely to encounter here on earth. In deep space, there is no shortage of H1 (in fact, most of the galaxy's mass may be H1) and even though this is a very rare transition, it occurs often enough that the vast clouds of H1 in the galaxy give out a constant hiss at this frequency. In fact, the 21 cm line was the first to be detected by radio astronomers, and we have been using that wavelength to map galactic structure ever since.

Because of its ubiquity, and its location in a region of the microwave spectrum where there is little interference from other natural transmissions and little quantum noise in radio receivers, it has also been a frequency favored by SETI researchers. The reasoning is that any technically advanced species would be familiar with this wavelength, and presumably, they would know anyone capable of building radio telescopes would be listening on it during the course of their astronomy research.

But why transmit at a frequency where most of the galaxy's hydrogen is emitting, potentially drowning out a faint signal? It seems more reasonable to transmit at some harmonic (twice or half) of the 21 cm line, providing that still is in the low-noise region of the radio background. Another strategy would be to transmit at some frequency formed by the product of 1420 MHz and some natural constant like pi or e. This could only be an artificially produced signal and would immediately be recognized as such. If I were doing active SETI, that's where I would transmit, and if I were doing passive SETI, that's where I'd listen. It always seemed to me a 21cm transmission would not be a good choice for SETI purposes.

As for a cometary origin for this signal, I'm not sure I'm convinced.
Is H1 found in the solar system at all? Would the amount carried aboard a comet (presumably from ionized water, methane or ammonia) be enough to be detectable? Granted, the comet is a lot closer than the neutral hydrogen clouds of the spiral arms, but there isn't much matter there. I don't know enough about this to do the numbers.