In microarray, there are 2 types of probes. I wonder why we need to use mismatch probe to know the gene expression? In addition, considering mutation, the substitution nucleotide can be vary a lot for the whole sequence, it can be at any position, so how can we compare when the mismatch probe is just a case or we just consider the perfect probes and mismatch probe is represented for "no expression" status? Thank you
Hi Phang, Are you asking about Affymetrix arrays specifically? I think other platforms rarely use mismatch probes. Historically, when developing technology for transcriptomics, affymetrix found that it is not that straightforward, there was a problem with background noise. Affy went with control mismatch probes to subtract that noise to some extent from the signal. They use multiple probes as well to combat the same problem, I believe. In Affy mismatch probes - the mismatch right in the middle of probe, position 13 from both ends of 25bp olig. They developed complicated data processing protocol, which worked to some extent but... Later a number of groups came out with different approaches based entirely on perfect match signals. That approach(es) works better. That my one-penny worth answer.
Thank you a lot for your answer. You are right, I mean about Affymetrix. Could you please explain more for me about the "noise"? Because my English, academic vocabulary is not good. When I listen to some video, they also mentioned about the noise and then they talk about the mismatch. I think that is exactly you talk about. But actually, I do not understand what is "the noise" in this case. (I do not think it is related to voice) Is it there are some mutation of the gene can effect to our result? I mean, if there is a mutation not influence on the gene expression, but our probe can not detect this mutation, then, a big number of gene expression will be neglect, and the result is not right anymore. Is it right? Or there are something I misunderstand? Thank you so much
Hi Phang, Noise in this case means hybridization to specific probe ( let say GAPDH) of other than GAPDH transcripts. So if you remove GAPDH mRNA from total mRNA prep, there will be still some signal for GAPDH probe(s) coming from that pool of mRNA ( when there is NO mRNA for GAPDH). That refers as a noise in the system. It will complicate the analysis. So that noise should be subtracted out. Affymetrix decided to have a probe close to specific one by having just one mismatch in it, with idea that it will have similar level of non-specific hybridization as a specific probe. And that subtracting that signal from perfect much signal they will eliminate the "noise". That is not that simple, but it worked in most of the cases ( genes and probes) for them in the first version of data processing soft. But again, latter people came up with better way processing data from affymetrix - by excluding mismatched probes completely, just taking in consideration only sequence of perfect probes and predicting how they will perform on the bases of that probe sequence. Hope it helps. Ask me more, if you need. :)
Thanh you so much, now I understand better. You are right that we have more modern way (than Affymetrix). Maybe I will look for the new one instead of Affymetrix.
You could still use Affy, but for data processing to use there affiliate Data-soft they themself recommend. Good luck!
- Badges
- Science topic
- Similar topics
- Biological Science
- Genetics
- Gene Expression