Novice: Statistical Analysis with few samples and missing data

[random_noob]

Hello!

Recently a colleague reached out asking for help wondering if there was any kind of statistic measure that could be used to evaluate some results.

The results are given in the table below.


NA - Not Available

There are 9 observations, but only three have a category assigned because there's no more data available (a poor number of samples was provided).

With this in mind, is there any kind of statistical analysis possible on this data, from which we could extract some information on the relationship between the Log2 Fold Change and the Category or should we just give up?

Another colleague suggested calculating the p-value and using bootstrapping/permutation tests, but I don't see how that could be possible in this case.

Any suggestion is welcome!

Thank you advance for your help.

 

[hlsmith]

So the problem is that the NA's should be actual categories - but you don't have this information? So you only know the actual category for 3 out of the 9 observations?

 

[random_noob]

So the problem is that the NA's should be actual categories - but you don't have this information? So you only know the actual category for 3 out of the 9 observations?

Exactly!

 

[hlsmith]

Yeah, I don't see much you can do. But I am on the outside looking in. You have three values to compare with only one observation per group. So you have no measure of dispersion on estimates for the groups (standard deviation or standard error).

Not sure if possible, but I would recommend starting over.

 

[random_noob]

Yeah, I don't see much you can do. But I am on the outside looking in. You have three values to compare with only one observation per group. So you have no measure of dispersion on estimates for the groups (standard deviation or standard error).

Not sure if possible, but I would recommend starting over.

I see. Unfortunately, I don't think starting over is a possibility. Nonetheless thank you a lot for your fast reply and contribution!