A DNA molecule could be single-stranded, double-stranded, triplex or quadruplex in structure. It could be linear or circular. It could exist in several alternate structural forms. It is not possible to define it in terms of a fixed length as it could be a few nucleotides to the size that is packaged in entire length of a longest eukaryotic chromosome.

We can measure single strand DNA and doubel strand DNA by measuring at 280 nm. it has a high optical absorption. it is a single strand DNA and if it has a low absorption it would be doubel strands DNA.

I agree with Jon. Technically, dsDNA would be two molecules. During replication, there could be thousands, until DNA ligase has done it's job. I guess anything from a dinucleotide upwards would count as a DNA molecule, as DNA is a polymer.

I think the crucial question here, is why do you want to know? Are you really trying to ask how many MOLECULES of DNA there are in a cell - why? - or is there a question behind the question? With the possible exception of qPCR, I find it hard to conceive of many scenarios where the number of molecules is important.

I agree with Jon ahd Richard (and Wikipedia), a molecule is defined by covalent bonds; a(n undamaged) double-stranded DNA chromosome is really a complex of two single-stranded molecules held together by a variety of non-covalent bonds and forces, and decorated with a plethora of proteins that form chromatin. As Richard points out, one chromosome undergoing replication typically contains many DNA molecules, with the leading strands replicated in longer molecules and the lagging strand replicated in shorter molecules that are eventually ligated together. DNA damage can covalently link the two strands together, and in that case a chromosome can become a single molecule. Likewise, single-stranded and double-stranded breaks, either catalyzed by enzymes or due to DNA damage, can change the number of DNA molecules present in the cell. Due to these complexities, it's hard for me to see a situation where the number of DNA _molecules_ in a cell is very relevant. There are many other terms for specific cases: strand, chromosome, chromatid, chromatin, Holliday junction, Okazaki fragment, haploid, diploid, etc.

The most useful way to think of "how much" DNA there is in a cell, or a sample, depends on exactly what question you're trying to answer.

If I may wax a bit philosophical, in science, the definition of an entity is useful if:

1. It facilitates expression of ideas concerning the entity. For example, if we did not define "energy" as encompassing both kinetic energy and heat, then the conservation law would become very difficult to express in a useful way.

2. Most people (within some group) can agree on it.

In science, it is often useful to define items operationally. For example, if you ask a physicist for the radius of an electron, she might reply, "First, tell me how you want me to measure it." An atomic physicist might say that it is 2.8E-15 m; a high-energy particle physicist might say it is zero.

If you want to count molecules, one way would be to apply the ideal gas law, n=PV/RT. If we consider a thought experiment, we might imagine putting DNA in a chamber with temperature & pressure such that it behaves as a gas. If the temperature is below the melting temperature (say, 60 deg C), then the double helix will count as one molecule; if above, the single strand will count as one molecule.

So, as Max has said, it depends on the question you're trying to answer.

The "it" I said depended on the exact question you're trying to answer was which term should be used to ask the question (which depends on how you intend to use the answer), not the meaning of the term "molecule".

There are good examples of concepts which are only defined/definable in terms of the experiment to measure it, and you gave a good such example: the radius of an electron. In the standard model of particle physics, an electron is not modeled as a physical object with a well-defined physical radius; thus without coming to an agreement on the meaning of the term, that particular question cannot be answered. The physicist is essentially saying "Ask me a question that is better defined!", while I was essentially saying "Will the answer to that exact question be useful to you, or would you like to ask a more useful question instead?"

In this case, the original question was about how many molecules of DNA there are in a cell. Since DNA molecules are never simultaneously in the gas phase and in a cell, I don't see _defining_ DNA molecules by a gas-phase experiment as particularly relevant to this question.

A single molecule of DNA would be represented as a single strand as described above by Jon. However, the functional unit of DNA will be either single-stranded or double-stranded depending on the organism and how it maintains its DNA in a native form. While the functional DNA molecule in humans is double-stranded, in some viruses it is single-stranded. So in the original question, each chromosome (prior to DNA replication) is composed of 1 functional double-stranded DNA that is composed of 2 DNA molecules. After DNA replication, each chromosome is composed of 2 sister chromatids and each chromatid is composed of 1 functional double-stranded DNA that is composed of 2 DNA molecules. The sister chromatids are not part of a single DNA molecule because they are not covalently joined together, but rather are tightly associated with each other at the centromeric region. As for the telomere, there is not a covalent joining of the two DNA strands at the telomere. The telomere is a single-stranded region at the 5' end of each DNA strand in the double-stranded molecule that serves to protect the 5' ends from degradation.

As per your comment about the number of human chromosomes, there are 46 human chromosomes, with each chromosome composed of a double-stranded DNA molecule (it is best when describing a DNA molecule stating whether it is double-stranded or single-stranded to clarify the functional nature of the DNA). However, because humans are diploid organisms, we contain 2 copies of each chromosome (except for the sex chromosomes, where females have 2 X chromosomes and males have 1 X and 1 Y chromosome). We use the nomenclature of 2n to refer to a diploid, which means that humans have 2n = 46 chromosomes or n=23, which refers to 23 unique chromosomes (22 different autosomal chromosomes and 1 pair of sex chromosomes). Thus, in females, there are 23 different chromosomes in a diploid state (2 copies of each unique chromosome to form a homologous pair of chromosomes) for a total of 46 chromosomes, each composed of one double-stranded DNA molecule. There is no physical connection between each homologous pair of chromosomes, so they represent two chromosomes.

I agree with those who point to the most appropriate response to the question. Possibly I deviate from the subject of the question, but I think David has introduced an interesting term, i.e. the number of copies of DNA. Depending on what the object of the question is, it may be useful to work based on the number of DNA copies existing in a sample. If it is to do any research on one or more specific regions of DNA (using PCR-based techniques, among others), knowing the amount of DNA or the number of copies, as well as the state of degradation of the DNA, can have more interest than the number of molecules itself.

I consider each single DNA strand to be a molecule. Including hydrogen bonds and other non-covalent interactions in a "definition of a molecule" would not only give you double stranded DNA, but would include all the proteins that constitute chromatin. The single strand of DNA is also the functional unit as DNA is replicated or transcribed by polymerases that "read" a single strand. Consider a water molecule. The water molecule is hydrogen bonded to other water molecules, but we don't consider the volume of water and all the molecules it contains to be the molecule.

Single-stranded DNA, since it is covalently bound, is one molecule. Double-stranded DNA has two molecules connected through hydrogen bonds and hydrophobic interactions.

My opinion is that you must first state if you are speaking of DNA as a chemical compound or as a biological object, because the two are conceptually different. As pointed out by many commenters, a chromosome (or a chromatid) may happen to get some accident during its "life" (i.e. between two replication events). So, the functional (biological) point of view may help to get a provisional answer, that is conceptually operative. Irrespective of its actual dymanic/metabolic state, a DNA double/single stranded molecule (depending on the genome structure) contains a set of instructions (= coded information) which can operate, by the assistance of external tools such as proteins and other biomolecules, to transmit/express the information. The point with DNA a a single functional entity is that all those instructions typically function as "cis-acting elements", as genetists call them; instructions are meaningful for sequences on the very same DNA molecule (some of them are meaningful only for the double-stranded entity - that's why I think a rigorous chemist's approach will fail to help). If you skip a few complex mechanisms that escape from the above statement of cis-acting elements, this can give you a definition of "single DNA molecule" in a living cell. You know, there are no universal rules in biology, you can always find exceptions.

Thank you everyone for your input.