- So replication is important part of experimental science. It's a process where some aspects of your experiment such as the samples within your group or the measurements being taken and in some cases even importantly the entire experiment itself, is treated exactly the same way and then the measurement is made.

- Replication is critical for understanding the truth and getting around the noise in our experiments. For example, if you wanted to study the yeast cell cycle and you only looked at one yeast cell dividing and measured one cell cycle. You probably won't be very confident in thinking that you figured out the true mean cell cycle length for all yeast. And you won't be confident that you had reduced any noise in your measurements. So it's really important to do replicates both within the experiment to look at more than one yeast cell and also to repeat the experiment to make sure that there wasn't any bias in the particular run of the experiment you did that you get the same results when you do it again.

- So replication comes in many forms, biological replicates, technical replicate and a repeat or independent experiment. Biological replicates are parallel measurements of biologically distinct samples. So there are a number of samples in each group which could be an experiment or control. An example of biological replicates are distinct plates of cells that are treated in the same way. So for example I may have taken an initial batch of HeLa cells. Divided it into three different plates and then performed the experiments on those three different plates.

- So you can use multiple biological replicates to correct for any random fluctuations in the data that might exist from sample to sample. So if there are very slight generic differences from sample to sample that affect its response to particular treatments by increasing the number of biological replicates then you can increase your confidence in what the true value is.

- So biological replicates test the intrinsic biological variability of a system. By determining a range of differences that might be observed when each sample is treated the same way. Now technical replicates are distinct repeated measurements of the same sample. So they're really testing the intrinsic variability of your experimental technic. An example of technical replicates are pipetting the same sample for example on the three different wells when running quantitative PCR reaction.

- Technical replication comes up a lot in biology because technical replicates are essentially multiples of the same biological material. So the biological material is not distinct genetically but you have repeated it's measurement a couple times. So lets say that you have genetic material that has come from one individual but you are assaying it multiple times. And each of those assays is identical to each other as much as you can ensure but because they were pipetted separately or they were in separate wells and a plate there is some variation between them. But the variation is not biological in nature because the biological material is the same. That's what makes them technical replicates.

- Technical and biological replicates are capturing different aspects or different sources of variation in data. And hence we treat them differently when we analyze data. It's important to keep tract of whether replicates were technical or biological because they will affect the statistical analysis downstream.

- So replicates serve as internal controls or checks on how the experiment was performed and the reproducibility of your technical expertise in conducting that experiment. So they're often taken to monitor the performance of that experiment and the reliability of the technique. However, replicates are not independent test of the hypothesis and so they cannot provide evidence for reproducibility of the main results which you need to put forth. For this you need to do a repeat experiment also called in many cases an independent experiment. And this is where you really begin the experiment at the very beginning and take it all the way to the end usually on a separate day. In some cases you may actually have to use independently prepared reagents which could be important variables in your experiment. And could be important for interpreting the results. So independent experiments are crucial for really determining whether that whole experiment is reproducible by you. Ultimately also whether it could be repeated by other laboratories as well.