Many general lab updates

For our nucleosome project, we have had some setbacks, most of which we have recovered from. As you remember from last time, our column had broken. That is still the case and we are still waiting for a replacement as it is on back-order. What I had failed to mention was that the sample itself had also had some issues. Namely, there was a white powder that had precipitated out of the sample and to the bottom of the tube. This was disturbing as I was relatively sure that this was all of my nucleosomes aggregating and falling to the bottom of my tube (something they are only supposed to do when I want them to). Since we were stuck without column, I thought I would investigate this issue.

It turns out that nearly all of my nucleosomes had aggregated. After some investigation, it turns out that the nucleosome sample was actually in 10mM Calcium Chloride (a +2 ion) instead of 1mM. This is an issue because nucleosomes aggregate at about 5mM of +2 ions. It is also a problem because the EDTA we add to stop the digestion is added to stop 1mM of CaCl2, not 10mM.

The first thing I needed to do was get the nucleosomes back in solution. We did this by dialyzing the nucleosome solution (and the precipitate) with 150mM NaCl (a +1 ion). This is done by putting the nucleosome solution in a bag that only lets ions and water pass through it, but not the nucleosomes. It turns out, if you add enough +1 ions, they will displace the +2 ions at which point the nucleosome should go back into solution. After a day of dialysis, we had significantly increased the amount of nucleosomes in our solution. After a weekend of dialysis, we had gotten back basically all of our nucleosomes. Now we had to make sure that they had been digested correctly (especially checking for over-digestion which shows up as smearing on our gel).

The results look like this. Now the furthest column to the right is our DNA ladder. This is used to let us know how long our DNA is. You match the known length of the ladder band to your sample. The bright band at the bottom is a 100 base pair DNA. Each step "up the ladder" is a 10 base pair step (110 base pair, 120 base pair, etc). The lane immediately to the left of the ladder is our sample. If you look and count carefully, you can see that it lies between the 140 base pair and 150 base pair "steps" of the ladder. This is perfect as the DNA should be 146 base pairs! Even better, there is no "smearing" to lower parts of the gel, which would indicate over digestion. You can see an example of this in a different sample that was run in the lane immediately to the left of our sample.

So as a wrap-up, we have the same amount of nucleosome we started with, all in solution and digested to the correct length. All that is left to do is to separate out any small amounts of double or triple or larger nucleosome arrays and we will be done. Since we are still waiting for our column to arrive, we may end up doing this through a sucrose gradient instead.

Not to be shown up, John is busy plowing through the ton of data that he has taken. Now that we have all of the data, we need to figure out why some of it looks great and some of it looks not-so-great. Things looked pretty bad at the beginning of this process, but each time I check in with him, he seems to have found another correction to make the data fall in line. If all goes well, we'll have it all whipped into shape by the end of this week.

July 27

Today I ran tests on the Lambda 25 to see if it is properly working, and it passed all the tests I could do without having to make any samples. I also created an instruction sheet for the machine with the basics so that anyone can now use it.

July 26th

Today I made and ran a second series of concentrated samples for the first series. The results came out somewhere in between the diluted series and the first concentrated series that I ran. This is promising, because the results should all be the same. It is possible that there is some sort of concentration dependent interference that is happening, which could be causing some of the differences. Today, I decided to start running concentrated trials of the second series, and started preparing the samples. This should be easier because there is much more of the second series left. Overall, I am finding that the eighth and tenth points are still uncharacteristically high, which leaves me to believe it is something in the actual sample preparation. This is one of the only explanations I can see left, because I have ruled out most of the possible problems on the testing end. One possible explanation is that not all of the Mg buffer solution was removed when the DNA precipitated.

July 23rd

Today, I wrote the second of two results reports, and updated all the graphs for the second series. It was Matt and Andrew's last days, so we had their final presentations today as well.

July 22nd

On Thursday, I began by analyzing the concentrated trial for the first series. It came out to be significantly lower than the eight diluted trials I had run previously. By lower, I mean the the normalized concentrations for the concentrated trial should have been the same as the diluted trials, but they were consistently below the diluted trials. This was kind of disturbing, and has prompted me to run more concentrated trials. I also updated my results report, and fixed and updated all of the graphs that I have been creating. I also consolidated and organized all of my data and graphs.

July 21st

Yesterday, I started by running the last trial of the second series. I also prepared a few more tests and ran them as well. The first was a new test for chlorine, which worked slightly better than the first. Part of the problem was that I didn't prepare the NaCl standard quite right, but some of the results still come out negative. I also ran some of the Chromatin that the professor and Travis have been working on, and found that it had higher calcium concentrations than they expected. I also designed and ran a highly concentrated series of the first series. This should help identify what is actually going on in the last few samples of each trial. They have been fairly random, but almost always are way above where I would think they should be. This is attributable to the fact that are very low concentrations of all of the elements. This makes sense, because less DNA is precipitating.

July 20th

On Tuesday, I continued to run the second series of samples, and I finished all but one of the eight trials. I also began to look at the spectrometer's ability to analyze chlorine in samples. This will be useful because it is possible that the cobalt hexamine ions are not completely dissociating. What this means, is that some of the cobalt ions we would expect to be +3 will actually be behaving as +2. Knowing the exact concentration of Cl would allow us to see if all of the ions were dissociating. Unfortunately, my first results came out very strange, with the chlorine concentrations going negative for a lot of the values.

July 21

Yesterday I spent the morning looking through the spectrometers manuals trying to figure out why the results we are getting are so weird. While doing and alignment i put our sample container into its slot and noticed that only a small amount of light was going through the sample compartment and the rest was just going over it. So the professor re-ajusted the holder and I ran another trial on matt and andrews red and purple solutions and got better results than the first time. I also re-ran the collaborators chromatin and got its absorbance to be 3 times what it was when i ran it the first time.

July 19th

Yesterday, I continued the testing on the second series of samples. I prepared three more trials, and tested four of them. I also began analyzing the first few trials that I have run. I've found that the ends of the series, near 45mM Mg have final concentrations that are very low, near 1-2 ppb, which is very hard to measure. While this results can still be used, it is possible that they are inaccurate. I also started correcting and adding more graphs to my report. I am adding some total charge and final concentration graphs. Due to the low concentrations, I am going to look at designing a concentrated trial today.

July 16th

Today I prepared the next three trials for the second series, and ran two of them. Everything is going well so far. One thing that I am considering is running a trial at a higher concentration. I think it would be interesting to see if only diluting the solution to 50X or 20X had any effect on the results. I think part of the reason that the results are going so wacky for the higher concentrations of initial magnesium is that the precipitated DNA concentration is so low. I think that running it at a higher concentration would help get accurate data for the higher points. Unfortunately, we don't have enough sample right now to do this several times, so I will have to plan it carefully to conserve sample. We also had our meeting updating everyone on progress, and also describing some of the work we still need to do.

July 15th

Yesterday, I ran the first trial of the next series. This series goes from 0 mM to 45 mM initial Mg concentration. As I looked at the results from this first trial, I was glad to see it closely resembled the appropriate part of the first set of trials, which went up to 22.5 mM. The last few points, towards the higher concentrations, and further trials should help see what is going on there. I also reviewed my results report for the first series, and updated all of the graphs. Unfortunately, we found an error in how I was calculating the errors in the measurements. Due to the way I uses excel spreadsheets repeatedly, this error was in all of the data I have analyzed so far. I was able to track down all of the repeated mistakes, and I believe my analysis is no more accurate.

Past few days

Basically in a nut shell we did another trial digestion using .1 microliters of nuclease instead of 1 microliter and we got the best results to date. We took our data and determined that we would get the best samples with a 20 minute digestion so we did a digestion on the rest of our chromatin. We set up our size separator in the cold room and as we were pouring in the chromatin the professor noticed that our tube had dried out making it useless, so we quickly poured our DNA out and lost a fair bit and then took the tube back to the lab to clean it out. However, during the cleaning we didn't use a Styrofoam sleeve to protect the glass and accidentally over tightened and broke our tube, so now we are at a stand still waiting for new materials. With my free time i have worked a lot on the website and did quite a fair bit of reading. Today I took the scatchard plot of the nucleosomes from a similar experiment to ours and used a program to scale and choose each of the points to figure out what their actual locations are and then turned it into a read-able graph that we will look at on Monday.

Great Successes and Amazing Failure

A lot has happened since I last posted. Let's start with the positive.

We have successfully run a very nice non-denaturing poly-acrylamide gel. The key was to increase the amount of sucrose in or loading solutions. This helped minimize diffusion of the samples before they started running. Another thing that helped was a mistake I made. In my haste to make the gel, I accidentally inserted a 10-well comb rather than a 15-well comb. However, this actually resulted in much nicer results.

We see exactly what we would like. The first column is our normal DNA ladder. The second is the undigested chromatin. As you can see, it is all at the top of the gel. Then we start digesting. As we go to the right, the time of digestion is increasing. As you can see, by the sixth column or so, the size doesn't change much for a few columns. This is the digestion being stopped right at the nucleosome core. After a bit of time, however, the digestion proceeds past this point in some of the nucleosomes, digesting the DNA that is around the core. That is why the gels get smeared towards the bottom for the last three columns. When we do our final digestion, we will make sure that we don't let it digest this long.

After this great success, we digested one of our samples and were going to pass it through our size-exclusion chromatography column. This column will separate our sample by size. This will allow us to collect all of the nucleosomes that are the exactly correct size, while ignoring those too big or anything that is too small (bits of DNA, etc.). The setup is shown to the right.

One thing about a size-exclusion column is that it must never dry out. The tricky thing is that you need it to almost dry out to add your sample so that your sample stays relatively concentrated in the gel. To make a long story short, I brought the column to the almost dry level and the closed the valve. However, the valve leaked slightly and after 15 minutes, when we loaded the sample, I noticed the gel had completely dried out.

At this point we needed to recover the sample (which we did right away). We then took the column back to the lab to recover the media (the white stuff in the column) so we could rehydrate and repack the column. During this process, I overtightened the column in the vise that was holding it. So now we are waiting on a replacement column and replacement media. It's funny how much a small leaky valve and 15 minutes can change things.

July 14th

Today, I had the Spectrometer running the entire day, and I made it through all 4 trials that I wanted to run. So far, I have found all of the results to be fairly consistent with my first set of four trials, which is good news. I think that the new calibration standard may have helped to lessen the extreme errors I was getting with the eighth and tenth points. All together, I now have 8 trials that I can include in my report of my results so far. This should give us a very good idea of the patterns of the ion concentrations. Tomorrow, I am going to start preparing the second series, which is also a magnesium series. The second series goes from 0 to 45 mM Mg. I will also being finishing up my report of my work so far, which I will probably be putting up on here. Travis and I have been putting together a website that summarizes the work that we are doing, which can be seen here: www.gburgsummerphysics.webs.com. This will be a good overall explanation of the theories and processes involved.

July 13th

Yesterday, I analyzed the final of four initial sample tests that I have been running. It came out to be similar to the other three trails, which is good for consistency. I will now be moving on to the next set of trials for this first set of samples, which I expect to be even more accurate. Using a better calibration standard should make this set of data more reliable. I also tested my new calibration standards, and found that they worked quite well. After diluting each one to the proper concentration, I am left with about 80 test tubes just waiting to be tested, which I should hopefully be able to finish tomorrow.

July 12th

Yesterday morning, I read two different articles. One was about DNA stretching, and the other was about incomplete ion dissociation. The DNA stretching article was not as relevant as I had hoped, but it was still interesting. It was about a statistical mechanical model for the condensing of DNA under tension. It was based on data similar to that which I am collecting, on the condensing of DNA in the presence of Cohex. It gave approximate values for the tension present when DNA collapses. The second article provided another theory as to why the DNA would condense and the resolubize under increasing amounts of Cohex. It suggests that as the concentration increases, not all of the Cl- leaves the Cohex, so there are increasing concentrations of Cohex+2 instead of the normal +3. This would mean that the Cohex is actually out-competing itself, causing the DNA to return to solution. It is possible that our results could be supporting this theory, although it is hard to tell without the Cl- concentration. This is unfortunately impossible to get because the solution was flooded with NaCl to resolubize the DNA. In the afternoon, the nitrogen finally got here, and I was able to resume testing. The first thing that I tested was the fourth of four initial runs that I did on the real sample.

July 9

On Friday I spent the day preparing another trial digestion this time using .1 micro liters of nuclease instead of 1 micro liter. Because our previous gels showed our digestion ending very quickly we decided that we were adding too much nuclease making the whole process end after around 5 minutes. So I did another 50 minute digestion and made a 10X dilution. I also did a digestion on the collaborators nucleosomes using the same method. For these samples we also added a lot more sucrose than before also in higher concentration. The professor hopes that this will make our gels much more readable then they have been.

July 9th

Today, I spent pretty much all day making new samples to test. I was anticipating getting the new nitrogen tank today, which would let me run some of these samples, but it hasn't come yet. I have two new calibration standards, which are both designed to span the exact range I'll be working with. They are both the same thing, but prepared twice just to make sure that sample preparation doesn't have anything to do with the results. I also prepared these calibration standards to be tested for accuracy as well, by separating them into additional tubes. I also diluted the DNA samples into 4 different sets, each of which are 100X dilutions. Today, the professor gave me a paper that has possible applications to our project. The paper suggest that some of the Cohex did not completely dissociate in the original solution, making some of them +2 ions. This would help explain why our results are significantly higher than the existing PB theory would suggest.

July 8th

Yesterday, I worked on analyzing the data in a few more different ways. I separated the data so that it showed just the average Mg, Co and P concentration. This helped me to explain why two of the points don't fit in my contributed charge graphs. This is a combined effect from the Mg and Co values being higher and the overall DNA (P) concentration was lower. I also started to design and make the next huge set of samples that I will be testing when we get more nitrogen.

And more DNA gels

Well, we keep getting better at least. This digestion showed a much better digestion. As you can see, the 0 minute digestion showed a smear of DNA, while all lanes that were digested for 5 minutes or more showed a nice single band that can be believed to be around 146bp. As can be seen we have a few issues.

First, the digestion is supposed to take around 30 minutes. This is helpful because it means that you can make sure that you are just digesting to 146bp and not overdigesting. Ours is taking 5 minutes. In our next digestion, we'll try decreasing the amount of nuclease (the thing that "eats" the DNA) by a factor of 10. Hopefully we'll then see a little more of a progression and less of a step function.

The second issue is that the gel still doesn't look great. The first issue is the ladder. We obviously didn't run long enough: Our 100bp standard only went about 1/2-way down the gel. However, conditions were controlled enough the last time that we should be able to have a nearly-perfectly timed run.

The other issue with the gel is that the DNA in all but the ladder lane is quite diffuse. It would be much easier to read if they looked like lines, similar to the ladder in previous gels. My theory is that this is due to the sample not being dense enough when loaded and therefore diffusing before it starts running in the gel. An easy fix for this will be to increase the amount of sucrose before loading these samples.

Finally, a problem with the gels that I'm not showing you is that I messed up the loading of the other gel. This is actually Travis's gel. Mine didn't have a ladder! Well, I'll have to do better next time.

Things with John's project are going quite well as you can see by his posts. We should have nitrogen tomorrow which will allow him to run more samples. After that starts up, we should be on our way to having all of the DNA aggregation samples done by the end of next week if everything goes well.

July 7th

Yesterday I spent most of the day reformatting the data that I have already collected. Instead of comparing the results across the different trials that I have run, I compared the results across the different collection methods the spectrometer uses. This gave me three more sets to compare, the Radial view and the two Axial views. This took a lot of manipulating the data, because everything I have done prior has been set up to compare different sets. Fortunately though, the end result was that I confirmed all three methods are equally valid. This means that I can keep using all three sets of data without fear of one of them being massively inaccurate.

July 7

Today we poured and ran our DNA gels.. the professor put in the concentrated samples and I put in our 10X dilution samples. To get the samples in the gel we had to take a pipette and slowly inject the solution into the wells in the DNA gels... After a while this gets difficult because you really have no room to mess up because if you go side to side you can mess up the wells and if you go to far down you can puncture the gel. But I finally got the hang of it and after some time we were able start our gel at 10V. The professor decided to use a power supply from the physics department because last time when we used the power supply that is in the lab it shut down during the night and messed up our results.

July 6

Today I ran another timed digestion going up to the usual 50 minutes. I created a 10x dilution of each sample and also made a sample of the old nucleosomes we had in the freezer by adding just enough to make the concentrations of the new nucleosomes samples and the old sample equal. tomorrow we are going to run the DNA gel which will probably run overnight.

July 6th

Unfortunately, the spectrometer ran out of nitrogen over the long weekend and we won't be able to replace it for a few days. This means I have to put off testing for a few days, but it also gives me time to look at the results that I have so far and plan the next set of tests. Today I worked on finding the uncertainties for all of my data and seeing how they compare to the theoretical Poisson-Boltzmann prediction. A few things bother me about the final fit that I came up with, which is seen above. First of all, PB theory would suggest that there should never be more than 1 total charge being compensated for, which there often is. Second, the eighth and tenth data points are disturbingly large. Third, the error bars for only the eighth and tenth points can be seen on the graph, meaning they are by far the largest errors. Fourth, the PB predicted curves shown on the graph as "Theoretical" had to be translated to fit that well on the graph. The Cohex curve had to be multiplied by a factor of 1.25, and the Mg curve had to be multiplied by a factor of 1.6. This means that my results are systematically higher than they "should" theoretically be. Also, I believe that once the new nitrogen comes and some maintenance is done to machine, combined with the new calibration standards I will make tomorrow, the results be more precise. This should help identify where the problems are coming from. Also, I should say that the third trial I ran, which is identical to the fourth I have yet to run, was prepared slightly differently and has already shown to be more accurate.

July 2nd

I spent most of today analyzing and writing up my results from the past few days. Some good news, is that the data sets are fairly consistent with each other. I graphed the charge that was contributed by both ions to compensating for the DNAs negative charge, and it actually appears to have a somewhat linear relationship. There are two points that do not fit the linear fit, the 8th point and the 10th. The 8th is somewhat disturbing, because it appears to be random. I originally thought I might have prepared the solution poorly, but now that I have run another differently prepared solution, I find that hard to believe. Who knows though, it is very possible it should be like that. The graph on the left is the average of the 3 data series that I have collected so far. It is also a little strange that there are often more than 1 total charge contributed by the sum of the Mg and Cohex, because DNA only has 1 negative charge per phosphate. Because the Mg and Cohex are normalized with the P concentration, there should never be more that 1 charge total. This suggests either some error, or some sort of overcharging. I am nearly done with the analysis paper I have been writing, and will most likely post it on here when it is finished. Next week I will be finishing analyzing this data, preparing the new calibration standard, and running as much more sample as possible.

July 1-2

Thursday was a pretty relaxed day as we waited for our DNA gels that shutdown over night to finish up. While we waited I read an paper on how Magnesium affects chromatin but it was a little dense and hard to understand so I'm still making my way through it. When we finally got our DNA gels and pictures of them we found that we had let them run too long and that the data we were looking for had run off the gels. But we had better looking gel than our first try so we are narrowing down our results and should get it the next time. Today I re-tested the old nucleosome samples to determine which one that we had was the best sample compared to our new DNA. after some confusion with the results I finally got the data and determined how much nucleosomes per gram are in the solution and how much we need in order to have a good comparison with our new DNA for the next time we run a gel.

New Gels, New Challenges

Note: I realize these posts may be confusing to those outside of biology/biochemistry fields. In general, gel electrophoresis separates molecules by size: In this case, biggest are at the top and smallest are at the bottom. For more, look at this great animated introduction.


As you remember from my last post, our last gel ran quite well, but the results showed that we had failed to digest our inter-nucleosome DNA. This time, we modified our digestion process (and by modify, I mean went back to the process I had been using before I decided to "improve" my process).

This time however, we have another issue. The gel itself gave us problems. There were a couple issues layered on top of one another, but the main one was that our power supply that runs the gel decided to quit some time in the middle of the night. This had two effects, one was that the DNA had time to diffuse and migrate in any old direction it wanted to during the time that the power supply was off. This causes a broadening of the bands which I think we are seeing here. Another issue was that it forced us, due to time considerations, to ramp up the voltage to very high values which I think is what caused the intense upside-down V-like shapes to the bottom of our gels. Finally, we ran the gel a little too long losing our 100bp marker. In the previous post, you could see the 100bp and 330bp markers, but I am pretty sure that the marker you see in the first lane is simply the 330bp marker. This makes it impossible to determine as, even though some of the middle markers are resolved, we have no way of counting up from the markers (each sub marker is 10bp).

As for the digestion itself, things look pretty good. We got rid of our smear of DNA and it is obvious that the DNA is migrating to approximately the correct place. There are only two issues I am worried about. One is that aside from the zero time point (the second lane), most of the lanes look the same. In other words, it seems as if we are completely digested by 5 minutes.

The second issue is that the lanes that do look different are completely random. The corresponding times are 20, 40, and 50 minutes. By 50 minutes, we shouldn't see anything above that band of brightness on the bottom. The fact that it is random seems to indicate a possible pipetting error. However, I ran and loaded two separate gels, so the pipetting error would have had to have come earlier. We'll try going over our process to make sure that this isn't an issue.

Other than that, it's back to do the exact same thing again. This time we'll be using a different power supply so we don't run into the same problems.

July 1st

Yesterday I began analyzing my results from the previous day's testing. The results seem to show a clear pattern, for as more Mg is added to the initial solution, more Mg is precipitated, suggesting it out competes the Cohex. In general though, as more Mg was added, less DNA is found in solution, suggesting that Mg does not precipitate the DNA nearly as well. I also began working on a report that shows my process and an overview of my results so far. I also prepared another set of samples that should be exactly the same as the ones I ran the last few days. Running these samples again should give me a data set large enough to get some good statistics for my results.

June 30

Today I spent the day making new timed digestion samples that we could run in the DNA gels. After I finished the final bath that the samples take in 50 celcius I created 10X solutions with each of the time trials because last time we made a gel it was very bright which is an indication that there is too much DNA. We ran the DNA gels overnight and tomorrow we are going to dye them and see our results