[spunkets]

" For instance, as applied to molecular biology, viruses can be seen as noise."

It can be, but then errors are introduced into one's understanding of reality. Physical noise is represented by well defined functions and no viral code is equivalent to a noise function. A viral code is represented by a function that generates a range of unique messages which are themselves subject to noise.

[Alamo-Girl]

Indeed, physical noise (e.g. thermal) is always present and in its worst form according to Shannon’s model has a Gaussian distribution.

But there is also a form of noise which is a coherent message or message fragment that is broadcast (non-autonomous to the receiver) or that bleeds into the channel.

For instance, you might be watching a television program and suddenly also be picking up a radio transmission. As the "receiver" the effect of the noise may be good or bad.

This phenomenon (autonomous v non-autonomous) has been posited as a pathway for mutation in biological systems and also a solution for the rise of syntax in abiogenesis (e.g. Rocha.)

You might find the peculiar applicability of Shannon's channel capacity to molecular biology to be interesting: Claude Shannon: Biologist