1 is the lonliest number

So today I am working on analyzing the seven different sets of data we took yesterday. I have finished the first two and am waiting for the third data set to finish compiling. Basically what I do is open up one of the videos we made of the dna/protien interactions and skim through it to find the clearest picture of the DNA. Then i use this matlab program called suivi that Rachid wrote. It takes the input parameters of the screenshot the dna is on as well as the total number of frames Then it opens up a picture of the frame with the dna andyou pick all the points you think are actually dna and not just dust. You can distinguish between the two because dna is brighter and lasts longer. So i click through the frames of the video and make sure that the points i am selecting last longer than 1 or 2 seconds. I normally pick around 30 points depending on the quality of the trial.

The program goes through each frame of the film,(generally around 2500) searching for high intensity points in the vicinity of the xy coordinates of the dna you picked. high intensit points hopefully correspond to proteins interacting with the dna but are often something else. Then the program generates a massive matrix with all the concievable bits of information you could want about every dna/protein interaction. It also makes a graph for each dna point you selected graphing intensity vs time.

I look through each of these graphs individually. On each two lines are plotted, one corresponding to a large area around the dna points, and another that is nestled closer and thus is generally mpre accurate. I search for spikes in these graphs which would indicate that something bright was right on top of my dna point. for each spike i check the xy coordinates of the bright spot and its distance from the initial dna. If the xy coordinates are rounded to a number ending with .5000 then the point is not accurate and should be thrown out. If the distance fro, the dna is greater than 1.2, the point is also thrown out because the error would be too large. For any points that do make the cut, I keep trackof the x,y coordinates,distance from, the dna, and the number of slides its visible for. This last number is the most important because the average lifetime will indicate what mechanism the protiens are most likely using to repair the dna. Right now we tend to average1-4 good interactions for every 30 dna we track for 2500 slides. This is obviously less than optimal and so we are going to try to improve this statistic by changing how we prepare the solutions (i.e. editing the incubation periods and wash outnumber of times)

Today I met Arnaud and Josephine who are also both interning here but for much longer than me. I went to lunch with the, and theyre a hilarious bunch. I spoke in french as much as i could and they alternated between english and french so i got to practice speaking and understanding french; oh happiest of days!