While examination of the differential expression of non coding RNAs from blood samples or cell cultrues and animal models, how many times should we repeat the microarray analysis experiments ?
Is the repetition of the microarray analysis change according to the examples? For example, we should repeat the experiments at least 3 times on cell culture model to identify non codings expression profiles, is that certain information?
Hi Mesut Karahan
number of biological replicates depends mainly on biological diversity you'll have in your samples. for cell cultures, 3 replicates are recommended but we can easily think of bench troubles that can reduce group sizes and prohibit further analysis. Take 5 replicates with similar qualities (for cell culture, RIN higher of 8) and produced by same manner, it will ensure your analysis and raise your statistical confidence.
fred
Frederic Lepretre thank you very much for your informations.. What about the human samples? How many times should we repeat microarray analysis on human blood samples?
Hello Karahan,
In my perspective its better to take the replicates in order to get the better results for omics studies as with replicates you will be able to apply statistics and gain certain confidence.
Hi Mesut,
There are basically two things, 1. Technical replicates and 2. biological replicates. Taking more replicates will increase your statistical confidence over the data-set. Biological Replicates are one with diverse biological samples for same experiment. Technical replicates are done with the same samples doing same experiment several times.
You are here stating about technical replicate. 3 technical replicates are sufficient enough for microarray analysis on ncRNA. Though, you can have any number greater than that for better statistics.
Regards,
Pratyay.
Pratyay Sengupta thank very much for your comment.
I want to ask that are these two techniques ; biological and technical replicates can change according to sample types for example human blood samples how many times or cell culture?
Sincerely
Hi Mesut,
The answer to your question can be readily calculated.
Its a question of what level of confidence you want (need) from your data: 80%? 90%? 95%? 99% 99.99%??
There are several good (and free) resources that you use to determine how many replicates (how deep) you need to analyse to reach the specified level of confidence in your data. Have a look at the following: Article Power and sample size estimation in microarray studies
; http://powerandsamplesize.com/; http://bioconductor.org/packages/release/bioc/vignettes/sizepower/inst/doc/sizepower.pdf.
Mesut,
These two are the methods for performing your experiments. Technical replicates are nothing but performing the experiment on same sample multiple times. This will reduce your error due to the sample handling and give results.They vary less. On the other hand, biological replicates are generated from different samples. They reduce the conflicts in context of biology. It can vary and give better result (which is conceptually more strong).
Now there is no conventional way to determine the sample specific replicate numbers that we can consider. Replicates are made only for ensuring your result statistically. Definitely, taking more replicates will give you more assurance. But we should keep in mind about the expenses for experiment. Thus if you have your biological replicates (2 or 3) with you, then while performing the experiment you can take 3 technical replicates of each of the biological replicate. Then combine them to get results.
Regards,
Pratyay.
Mesut,
Please do not follow Pratyay's advice.
This advice is not good science.
It is approaches like this that ensure most of the scientific literature is not reproducible.
The language of science is mathematics (or statistics if you prefer). Please apply robust statistical methods to determine how much confidence you might have when you interpret your data. (It will never be 100%) The objective is not to get a 'better result'; it is to find the truth. Truth is only found in irrefutable facts.
Ignore bad science. Do the math!
I think if a particular study has been done for quite few times and you just want to look for the some of the facts than you can chose few replicates.In addition to this it is important to find the truth first and then consider the numbers.
Thanks David Ian Leavesley for correct me. Applying robust statistics is the only way. More times we perform the experiment, result will be more accurate with less false positives. And definitely finding the truth is our objective.
- Badges
- Science topic
- Similar topics
- Biological Science
- Biochemistry
- Biomolecules
- Nucleic Acids
- RNA
- Small RNA
- microRNA