This solves nothing.

The companies that produce DNA (like GenScript) can make a strand of arbitrary length using any bases. This includes the normal A T C and G, but could also include methylated version as needed.

Testing for methyl groups has been around for a while by using methyl-sensitive DNA cutting enzymes. This procedure would just help you determine which sites to add methyl groups when you are manufacturing your fake DNA. Anyone who has access to a lab (or company) that can make DNA can probably also test for methylation.

One just has to be a bit more careful when producing DNA for crime scenes, though the free market of underground DNA forgers should come in line with the technology soon enough.

A lot of creatures in the ocean do that, usually to startle or distract whatever is chasing them. That wasn't what was so special about it. Most luminous organisms in the ocean glow blue; these glow green, and are rare in that respect (at least in the ocean, see: http://en.wikipedia.org/wiki/Foxfire_(bioluminescence) ). The original publication ( http://www.sciencemag.org/cgi/content/abstract/325/5943/964 ) was only claiming that they found several new species in the deep sea that were unrelated to known organisms. Of course, they have this media-friendly property that they release glowing 'bombs'.

Also, yes, whoever sees it first gets to name it - whatever they want. Obviously this can lead to much silliness.

They mention in the article they are examining 11 different genes out of over twenty thousand in the whole genome. Obviously they are not concerned with what those genes actually do, but only that certain variants may correlate with improved performance sometimes. The best estimates of heritability of intelligence are somewhere from 30% to 50% genetic, meaning at least half of the variation in intelligence is due to environmental factors or multiple gene interactions.

It's not about global evolution as about treatment effectiveness. If your target protein is rapidly mutating, then selecting against it is likely to either down-regulate that protein or select for other mutations where the antibody binds to reduce binding. Supposing that only 1 in 100,000 cells survive due to that mutation, then you could say that it is still 99.999% effective. But the one surviving cell would be the mutant that proliferates and returns months to years later. It costs millions to develop these treatments and get them through clinical trials. Having it only work some of the time just doesn't seem good enough.

Won't this only provide selective pressure for those mutant cells to make another variant of Mucin1? That is exactly how aggressive cancers form in the first place.

My guess is that anything Merck would find useful (patentable) they would patent, and leave the rest for this open source project. GSK is doing a similar practice. Part of the motivation is to encourage development of drugs for neglected diseases, that is, diseases where the market is small and big companies like Merck and GSK have no desire to drop the millions to develop the drug that wouldn't help that many people, or would only help poor people.

There are multiple epitopes, or targets, on any viral particle. The first b-cell to multiply and produce antibodies is whatever one sticks first, which is probably the most obvious target and also likely the epitope that mutates the most. A vaccine like this would work but it would have to be designed in a way such that only that one epitope is available for the immune system to generate antibodies. Maybe binding the target protein to nanoparticles and injecting it would work.