I am quantifying spots on an autoradiography scan, so subtracting the background pixels to get the intensity values but when nothing binds, it's the same as the background and since that's an average. It's sometimes negative. This is specifically more pronounce for the lowest concentration. Should I omit the lowest concentration value as it is beyond the detection limit and include a bit higher concentration to generate the calibration plot? Also if I decide to include it on the calibration plot, should I use negative value or consider it as zero?
As I understand your problem, you have a single value for the background measurement, and sometimes when this single average measurement is higher than your spots of interest. A detection limit is generally defined as the point where the signal is some number of times the standard deviation of the background noise level. If you only have a single average background level, you cannot determine its standard deviation, therefore you cannot determine a detection limit. The answer here is to measure the background in a number of locations on the autoradiograph and then determine the background and its standard deviation. With that information, you can then determine whether you have detected something or not, based upon your definition of your limit of detection. This is generally considered to be three times the standard deviation of the background, with the assumption that the noise is Gaussian, this gives a 99.7% probability that you have actually detected somethign.
As I understand your problem, you have a single value for the background measurement, and sometimes when this single average measurement is higher than your spots of interest. A detection limit is generally defined as the point where the signal is some number of times the standard deviation of the background noise level. If you only have a single average background level, you cannot determine its standard deviation, therefore you cannot determine a detection limit. The answer here is to measure the background in a number of locations on the autoradiograph and then determine the background and its standard deviation. With that information, you can then determine whether you have detected something or not, based upon your definition of your limit of detection. This is generally considered to be three times the standard deviation of the background, with the assumption that the noise is Gaussian, this gives a 99.7% probability that you have actually detected somethign.
Thank you, John Canham and Armando Hernandez. I have a second question. Can my measured intensity is still < 3*STD? what would be the reason for that? Systemic error in measurement? Thank you!
- Badges
- Science method
- Similar topics
- Chemistry
- General Biochemistry
- Biochemical Techniques
- Analytical Chemistry