Protocol Online logo
Top : New Forum Archives (2009-): : Bioinformatics and Biostatistics

A basic biostatistic discussion - (Oct/17/2009 )

Hi everybody!

We are having a basic discussion in my lab and I would like to hear another points of view. We're working with some cell lines and our discussion is about next:

If you have a 3 (or 6)-wells plate and you seed it: Is every well an independent sample?. or are they replicates?. In another words: How do you count the number of experiments? By wells or by plate?

I say that if you load onto a gel three samples got from a cell lysate (from a single well) you'll get a triplicate. So, for me, a 6-wells plate should give us 6 samples (probalbly indepents samples). Anyway, this is not how I proceed with my experiments. I consider that a 6-well plate it's just a unique sample (I count my experiments by plate).

What do you think about?

Thanks a lot!!

-interleukino-

interleukino on Oct 17 2009, 02:14 AM said:

Hi everybody!

We are having a basic discussion in my lab and I would like to hear another points of view. We're working with some cell lines and our discussion is about next:

If you have a 3 (or 6)-wells plate and you seed it: Is every well an independent sample?. or are they replicates?. In another words: How do you count the number of experiments? By wells or by plate?

I say that if you load onto a gel three samples got from a cell lysate (from a single well) you'll get a triplicate. So, for me, a 6-wells plate should give us 6 samples (probalbly indepents samples). Anyway, this is not how I proceed with my experiments. I consider that a 6-well plate it's just a unique sample (I count my experiments by plate).

What do you think about?

Thanks a lot!!



We do gentamicin exclusion assays, and we treat each well as a replicate, and repeat the experiment 3x to get our triplicate experiments.

So, our plates set up as follows:

-macrophages harvested from animal
-macs seeded into 24 well plate
-use 3 wells per sample per time point - have 4 bacterial strains and 2 time points, so that fills up one 24 well plate

This constitutes one experimental unit. To get multiple experimental units, we do multiple plates with macs harvested from different experimental animals.


We've done similar types of experiments using a cell line, and set it up the same way.

-fishdoc-

I don't do cell culture, but what your describing seems reminiscent of microarray work. If multiple samples to be measured originate from the same source, they are called "technical replicates". If they are the same thing (same cells, bacteria, etc.), but from different sources (cultures), they are called "biological replicates". Statistically, technical replicates are multiple measurements of the same culture, while biological replicates are comparative measures of different cultures. Both are necessary for statistical rigorousness.

-HomeBrew-

Thanks Homebrew!

That's a good point! And I guess that's why our discussion begin: difference between tech or biological replicates.

Thanks!!

-interleukino-

Again, using my microarray point of reference -- biological replicates are considered different experiments, while technical replicates are considered different measurements of the same experiment. So, the number of "experiments" you have equals the number of biological replicates you have...

-HomeBrew-

I've been reading about technical replicates, and some papers describe the use of 6-wells plates as a unique "biological" replicate (the same way I use to proceed). But I'm still in doubt, because every well should be useful to measure "biological variability" (I mean, every lisate from a single well). In fact, every well will suffer differents inhibition levels (eg in siRNA experiments) and other variables. I know this is a kind of "philosofical" discussion with a great relevance (and supose it belongs to experimental design) but I didn't see any clear definition in editorial policies of journals. I should to read a lot!! Thanks!

-interleukino-

Again, I don't do cell culture work, but in my opinion loading multiple samples on a gel from the *same well* are clearly technical replicates -- repeated measurements of the same experiment (sample).

Whether or not each well qualifies as a different biological replicate (experiment) depends on what you are doing in each well.

-HomeBrew-

HomeBrew on Oct 18 2009, 07:50 PM said:

Again, I don't do cell culture work, but in my opinion loading multiple samples on a gel from the *same well* are clearly technical replicates -- repeated measurements of the same experiment (sample).

Whether or not each well qualifies as a different biological replicate (experiment) depends on what you are doing in each well.



i agree ;)

-Pradeep Iyer-

hello guyes,

 

anyone can help to solve these questions, please 

 

The distribution of blood glucose levels in a large population of people of the same age and gender is roughly normally distributed.

For 15 year old boys the mean is 5.5 mmol/L and the standard deviation is 1.1 mmol/L.

 

Note: The following questions (part (a) to part (d)) relate to the this information and the values presented above have been altered for this assignment and may not truly represent a population.

 

(a) Levels above 9mmol/L may require medical attention. What percentage of 15 year old boys would be expected to have more than 9mmol/L of blood glucose?

 

 

 

 

(b) What percentage of 15 year old boys would be expected to have blood glucose levels between (and including) 5 mmol/L and 6 mmol/L?

 

 

(c) Approximately 99% of the blood glucose levels of 15 year old boys fall between  mmol/L and  mmol/L.

 

(d) If as an alternative to measuring blood glucose levels a researcher decided to instead look at the diagnosis of diabetes in the wider population,

what kind of probability distribution would you expect the proportion of diabetics to follow?

 

Student's t

 

Uniform

 

Normal

 

Binomial

 

 

thanks 

-Qutaibha-