Computer Virus that infects real-world DNA?

 The computer code, written in C++, hosts the DNA sequence of M.mycoides
 JCVI-syn1.0. At runtime it acts as follows:

 1) Preparing the DNA sequence of M.mycoides JCVI-syn1.0 in the memory,
    (with slightly modified watermarks).
 2) Encoding own file-content in base32. The base32 code is then encoded in
    JCVI's DNA-encoded alphabet.
 3) This representation of its digital form is then copied to a
    watermark of the bacteria's genome in memory. With this, a fully
    functional bacterial DNA sequence including the digital code is
    generated.
 4) Next it searches for FASTA-files on the computer, which are text-based
    representations of DNA sequences, commonly used by many DNA sequence
    libraries.
 5) For each FASTA-file, it replaces the original DNA with the bacterial
    DNA containing the digital form of the computer code.

 The code has a classical self-replication mechanism as well, to eventually
 end up on a computer in a microbiology-laboratory with the ability of
 creating DNA out of digital genomes (such as laboratories by the JCVI).

Answer:

It is NOT impossible for a computer virus to infect real DNA because ‘artificial’ life was already created from synthetic DNA by Craig Institute. Proof.
That synthetic DNA coming from files that were generated in a computer.
So, if the virus happens to be in the right computer, at the right
time… But the chance is small.
However, in the next 10 years, that virus might be a real problem as
every year it is more and more easy to create ‘artificial’ life.

IMPORTANT UPDATE!

I have found this comment in the C++ code:

// Encode own file into Craig Venter’s DNA encoding language

It looks like this virus was designed expressly for Craig Venter! So
the chances for this virus to work (affect real-world DNA) are even more
realistic!!

Validity of the code

Less experienced people doubt the validity of the code. However, no proof was brought
to demonstrate that the code might fail (and why). Though, I haven’t
spend hours to read the code line by line and thoroughly analyze it, I
did pick a look at it and it seem sound and valid. The fact that the
author of the code targets a very specific organization (that “happens”
to generate DNA from computer files) tells us that he knows what he is
doing.

Purpose

As somebody already brought to our attention, there is no purpose in
creating this virus. But this is totally irrelevant to the original
question. Even if the DNA virus will destroy the host cell and even if
it won’t be able to replicate in real world (and it won’t) it is
irrelevant. If in the right computer, the computer virus will pass from
the virtual world to the real world, and this is OP’s question (I
think).

Conclusion

The question asked was: “can a computer virus could infect DNA”?. The question is
ambiguous. If it refers to the DNA of some living cell the answer is no.
HOWEVER, if the question refers to a human-created cell, the answer is
YES. Craing Venter creates DNA from computer (‘fasta’) files. If one of
those files happens to be modified or ‘infected’ by the computer virus
then the computer virus will have effect in real world (the DNA
synthesized by Venter will contain the code injected by the computer
virus).