Hello Cheng,

I think it's not to worry about your extra data. You can report results as a mean +/- standard deviation (n=5) for TFC and simultaneously put results with n=3 for the same , And mention n=3 for other assays results. its not wrong to have extra data. and you are preparing thesis then definitely it will be acceptable, even research papers accept such kind of kind of additional analysis.

All well wishes for your endeavor.

This is a very common problem with every researcher who  do not believe in staistical calculation during submission of research plans.  You must have a premonition of what should be the tolerance limits of the data that you wish to generate depending upon the accuracy that you wish to get. Secondly you must also expect certain amount of deviation to get a satisfactory result. This is not working with biased mind. Your experiments will generate lot of data but your tolerance value and expected deviation will filter out most of the data and give you what you want. However, if you wish to peep in to the other side of the results (so that you can discuss your results properly) then those unused data can be processed to give you another picture which may be something new or unacceptable..

Do not worry, at least you have replication. Replication is essential for obtaining reliable estimates of method performance characteristics such as precision and bias.

For your future endeavours, it could be interesting the attached guideline.