2 replies to this topic

[fungus9]

Hello,

I have been running a few plates on our real time machine this week and I have a question for you all. I always run my samples in triplicates and use the average of the 3 numbers. Sometimes I will get very consistant results (ex. Ct values of 25.15, 25.17, 25.19). However, sometimes one of those numbers are way off, like 25.15, 25.17, 27.90. Can I exclude the last number when this is the case?

I have heard that many people go by different "rules." I've heard that people exclude nothing and take the average. I've heard that some people exclude one number if it is off by 0.5 compared to the other two. What is the RULE??

Thank you so much!! I'm new at this.  

Edited by fungus9, 05 August 2009 - 06:42 PM.

[eldon]

fungus9, on Aug 5 2009, 06:40 PM, said:

Hello,

I have been running a few plates on our real time machine this week and I have a question for you all. I always run my samples in triplicates and use the average of the 3 numbers. Sometimes I will get very consistant results (ex. Ct values of 25.15, 25.17, 25.19). However, sometimes one of those numbers are way off, like 25.15, 25.17, 27.90. Can I exclude the last number when this is the case?

I have heard that many people go by different "rules." I've heard that people exclude nothing and take the average. I've heard that some people exclude one number if it is off by 0.5 compared to the other two. What is the RULE??

Thank you so much!! I'm new at this.  

excluding data results in other labs not being able to reproduce your results.

bad science.

[ivanbio]

This is straight from Nolan et al. (2006) "Quantification of mRNA using real-time RT-PCR" (Nature Protocols 1(3): 1559-1582), a reference worth reading:

"All replicates should be within 0.5 Ct of each other. At low Ct the tolerance should be lower than at high Ct. Above cycle 35 the variability will be greater and quantification may be unreliable"

In other words you cannot exclude the 27.90 reading, otherwise you would be manipulating your data as eldon eluded.

Having a single reading diverge by as much as 1 Ct, or more, is likely due to some problem in your experiment. It could be a pipetting issue, a sample issue, etc. Try to determine what you could be doing differently between the times when you get close to no difference in your Ct values to when you get this large divergence. In my experience this level of variation in Ct values could be due to pipetting technique. If you are not already reverse pipetting, give it a try.