[MUSIC] We're going to go a little bit into the essay for transformation that was used. And I decided to pick this page from McCarthy's notebook because it gives some kind of a more personal flavor. So basically what they do is they take amount of material. Which can be either diluted, or not diluted, or more diluted. And in this particular case, they used 0.5 microgram. 0.301 and then it's always the same, a dilution, a dilution, a dilution. And finally you see that you go from one-tenth to 1 to 10,000. So you have a 10 to the 3 scale in this experiment. And basically there's very little information because only the last two lines are meaningful. The last line says that if you use 0.1 whatever volume. You don't get turbidity and you don't get s-strains. The line above says that if you use three times more, you get transformation. You get s-bacteria. But this is not quantitative. Whether a tube has one transformation event or a million transformation events, at the end of the growth you have turbidity and smooth bacteria. Granted if you have one, you will reach the saturation, you will see the turbidity, later that he could have a million. But not by much. The difference between, if the bacteria grow every 20 minutes, a difference from 1 to a million is 20 generation. 20 times 20 is 400 minutes. So six hours. What you do when you do an experiment like this is you don't stay in your 37 degree culture room to look at the tubes and decide when they become turbid, no you don't do that. You put them in in the afternoon, you go home, you sleep, you have dinner, whatever. You come back the next morning, you have coffee, and then you go to look at your tubes. You have no idea whether the tubes are saturated at midnight, at 4 o'clock in the morning, or five minutes before you went into the room. So this is a very, very crude acid. And Mike Seabrook, which we will meet next time, was extremely reluctant because the acid was not going to take. It's not that he didn't believe the data. But he believed the answer was too crude to be used. Okay. So one of the typical way to analyze your material once you've verified your material is to use enzymes that will degrade various biological compounds. For instance, you can use a phosphatase from bone, or you can use an enzyme that destroys lipids, or enzyme that destroy proteins. Or serum, provided it's been treated properly. And you know that all of these have different enzymatic activity. For instance, in the pneumococcus autolysates, there is no phosphatase. And what they see is that with this very crude system, is that when they can degrade the DNA from high polymer to low polymer, they destroy the transforming principle. It's a correlation. Now, this correlation is useful but it's not very strong. And they're aware of that. They're totally aware of the fact that the correlation is not very strong. Now, DNA polymerization or DNA size. It's about the same, is something that you've all seen, ever since you were kids. All of you have seen high molecular weight DNA. If you ever had a cold, if you ever had a drop, running from the nose, and dripping all the way to the floor without leaving the nose, we've seen DNA. That's the viscosity of DNA. Now we're reaching one of the key elements of the paper, which is titration, the quantity. Now this quantity is subjective to the same criticism as the previous experiment because it's the same assay. If you use one microgram of DNA, in four cultures, you can see that all four cultures contain transformants. Can be one transformant, can be ten transformants, can be a million transformants. Very little information from this line. The line that has 10, this is micrograms, so this is 10 nanogram. This line tells you that still at this dilution, you go from 104 less than the first line, you still have all the tubes in which transformation occurred. Whether it occurred once or many times. And so only was that one nanogram you find no transformance. And the most important line is this one. With three nanograms, you have tubes without transformation like this one. No, to risk transformation like this one. Yes, no and yes. Now, how can you take this data and make it do something that is an acceptable number? And was absolutely unable to talk to Avery. I mean, they were coming from completely different cultures. And basically what wanted was to have a solid analysis of this data, which means you grow 100 cultures. And you count, how many are positive and how many are negative. And then you use a distribution to analyze the data. Imagine you go to a park where you have a chess board on the floor. There are many parts in the world where they have a chess board on the floor. A chessboard is 64 cases, 64 little squares. Now, imagine, you go to this place, and you take 64 coins of the same weight, of the same size, and you throw them one by one. What will be the result? Well, the result, a French physicist named told you that the result will be about 35% of the slots empty, 35% of the slots one coin and the rest, 30% of the stock, two or more coins. That's just simple probability. So here, if you were to count the number of cultures, you would say that there, 50% of the cultures are empty. And 50% of the cultures have transformance. So the empty is what we call the P0 class, and this is 0.5, 50%. From this you can calculate the mean number of events in this series. If you go to the chess board and you count 35 empty slots, I don't have to go to my calculator to tell you, you have an average of one coin per square. If you have 50% of the slots that are empty or the cultures that are empty. Well then your mean will be of the order of 4.7, 4.6. That's the mean number of event. So three nanogram of DNA gives you 0.6 transformance. Of course, 0.6 transformance is absurd. Either you have a transformance or you don't. That means that with 5 nanogram of DNA, you have an average of 1. So you need 5 nanogram of DNA to get one transforming. And from this, you can calculate the number of molecules that of DNA that are present in your preparation because you know the molecular weight and you can make the little calculation with those number. And if you do the calculation. You find out that the many, many, many molecules that have to be there to see a transformation event. It's an amazingly poor process. Poorly efficient process. Now if you make the same calculation and assume that there is in your DNA 0.1% contamination with a single protein of average size, you find that there are more molecules of the protein than molecules of the DNA. And so the criticism laid by [INAUDIBLE] And the other, about the contamination with the protein, is still a valid criticism. But in spite of all this, what they had was essentially the best they could do, except for this quantitative analysis, this mathematically, this is not real math this is simple calculus but this is the Poisson distribution.