Re: How does DNA compare with integrated circuit technology?

Allow me to indulge in a little background so that we'll have a basis for comparision:

There are 4 nucleotides that DNA codes for: adenine, guanine, cytosine, and thymine. These form the variety of combinations which (indirectly) create the proteins that make up all life on earth; however, some combinations are duplicates, some are dead ends (equivalent to "start", "stop", "skip", and "oops"), and some stretches of DNA code are "alien", that is, they are inserted into our DNA by other mechanisms, such as a virus. DNA is divided into "strands" which, for purposes of this discussion, do not normally interact. Each living thing on this planet has a fixed number of strands. For human beings, it is 46.

Transistors are current corrals, used to amplify, control, and generate an electric current. The number of transistors that can fit on a chip is theoretically limited to the distance that will keep the electrons from one circuit interacting with the electrons of another circuit.

Bearing that kind of waffling in mind, here are the answers to your questions:

Qestion 1) If one could take all of the matter in a human's DNA and (magically) convert it to silicon and related elements, then how many transistors could be made with the converted matter?
Answer: Human DNA is made up of 46 chromosomes. These can be thought of, in relation to the question, as equivalent to circuits. These circuits are more concerned with the upkeep of structure (the body) than with the regulation of energy. Transistors are not at all concerned with the upkeep of the structure of the machine in which they operate, only with the energy output of that machine. It is better to think of the human body, within the context of the question, as needing 46 circuits to create another machine like itself. Transistor replication would not make another machine, only another circuit.

Question 2) How many bits of information are needed to fully describe a human's DNA, or how many megabytes of hard drive space would be required to store the information contained in a human's DNA sequence?
Answer: Scientists have not yet decoded the complete human DNA sequence. For more information on their progress, look on the WEB under "human genome" to see "The Human Genome Project". Seeing the raw data for DNA decoding is not the same as seeing what a paticular stretch of DNA will do. That will take decades.

Question 3) Assuming Moore's law for integrated circuit densities holds out (ignoring theoretical IC geometry limits), how long before we can manufacture matter having the information density of human DNA?
Answer: We already have. Information density is not the problem. Think of DNA as the "program" for the human body. A computer program can generate quite a bit of information. For instance, the accounting files of "Mom and Pop's Travel Service" have a lot smaller information density than the account files of Chase bank.

Your subject question was "How does DNA compare with integrated circuit technology?" is interesting because of the research being done to have circuits design themselves. The "Artifical Intelligence" programming of the past few decades have evolved into a promising line of research. Look on the WEB under "evolution algorithm". Although I've never met the man, I don't get a dime for endorsing him, and there may be many others in his field, you might also look up "John Koza". Thier emphasis is as engineers and programmers, not as biologists. I believe that to be your background, also.

Mac Salfen
--------------
Admin note:
Mac Salfen later provided a more detailed answer:

DNA is an "apples and oranges" comparison to an integrated circuit. DNA is much more analogous to a computer program than it is to a circuit. Within that context, your question seems to be asking: "How big a computer program codes for a human being?" and "How much electronic storage does it take to store that program?"

First of all: DNA is wet chemistry, not microcircuitry. In our bodies, cellular DNA codes mainly for RNA, which codes mainly for proteins. Proteins are manufactured on ribosomes outside the nucleus of a cell, in the cellular cytoplasm. So, DNA is just a design program. The real work of DNA is in design, not in program execution (mainly done by RNA), nor in program output (mainly done by ribosomes).

Second: Your question about reducing the design program of DNA to an electronic equivalent is being done by the Human Genome Project, although the actual goal of the Project is different. Of course, the contents of the computer storage circuitry used by that project will eventually be turned over to labs doing the wet chemistry necessary to verify exactly what part of the DNA does what.

Third: Your question about "information density" doesn't go deep enough. DNA only uses 4 nucleotides for coding. If I give you the sequence ATGCATGTGGGAATT, I may be giving you an actual DNA coding sequence, but is it useful information? Like any computer program, it must be inserted into the "Compile and Run" environment of a machine that can interpret the code. For DNA, that environment is a living cell. Again, the Human Genome Project is attempting to discover the actual sequencing of the DNA code, and more importantly, what that sequence does.

Fourth: DNA codes differently in different animals, or plants for that matter. The genes (sequences of DNA that form a meaningful set of subprograms) that determine where the eyes are located in a human will not necessarily work for a potato.

Fifth: DNA is subject to change. Not only do we inherit part of our DNA from our mother and part from our father, but the DNA within us, and possibly the DNA that formed us, undergoes mutations which make it different. The rate of mutation is used as a yardstick to determine how long ago we diverged from similar animals, and as a yardstick to determine how long we have been in our environment (you can look up references to Darwin's Galapagos finches if you want more information).

Sixth: DNA is coiled. When a particular piece of DNA is needed, it uncoils. The apparent tight arrangement of DNA is only because most DNA it is not being used.

Seventh: Ready for some really rough measures? Uncoiled human DNA is about a yard long. Within this yard are roughly 15million nucleotides. We could represent that as follows: Standard ASCII gives a value of "65" for a capital "A". In binary, this is "01000001", using the eight characters of a "byte". If we assign each DNA nucleotide a single alphabetic letter (such as "adenine" being the letter "A"), and then assign each of those letters a computer information byte without any other underlying code, we have a crude basis for the answer to your question. O.K., so now finally, we can compare the design program of DNA to the information storage of a CD or disk drive. Fifteen million bytes (remembering that each nucleotide would take 1 byte to represent alphabetically) is 15000K (we'll conveniently forget for a moment that a kilobyte of storage is actually 1024 bytes), or 15 megabytes. Fifteen Meg of storage is well within the capacity of any modern computer.

Now tell me: As an engineer, can you fit 15M of storage onto a yard of filament (remember that human DNA stretched out is about a yard long)? I think if you consider the way that information is "wound" on a hard disk platter or CD, there is more than a yard of winding space.

I hope this explanation comes closer to answering your question. I hesitated originally to give you the information in my 7th point because it is pretty much a WAG for this "apples and oranges" comparison you asked about.