0
$\begingroup$

I have a dataset with 5 groups : 3 consists of patients with different cancer types, one consists of patients with benign tumour, another is a healthy control group. The proteins are measured in a way that they either have a non detected concentration level (intensity = 0) or some nonzero value, though these are not absolute quantification, so the differences in these values are rather abstract ( not meaningful mathematically). Sample size is rather small - <90 subjects overall.

What researches did is they performed PCA on the whole set of predictors (200+ proteins) - there are no distinctive patterns between the groups. However, when they use the subset of proteins (~50, based on literature review - which showed some association with cancer in previous studies) - there is some mild separation visible between cancer and other groups. First of all, the dataset is 'zero-inflated', that is there are many proteins which are simply not detected for most samples (intensity = 0).

Second, As there are no clear separation, how this situation can be interpreted?

Third, what would be other ways of performing statistical inference on the data (other than , maybe, Fisher test of nonzero vs zero table between cancer and non cancer samples )

Thanks in advance

Improve this question
$\endgroup$
3
  • $\begingroup$ How did you measure separation and what were your criteria for "clear" separation. Also, with 90 subjects in five groups, it's going to be hard to see (or test for) any differences that aren't pretty darn large. $\endgroup$ Commented Apr 25 at 13:50
  • $\begingroup$ yeah, they didn't measure separation, they just stated that since cancer samples tend to cluster together on one side on the PC plot, while benign and healthy samples are on other side - there is 'mild' separation. I am wondering what is more suitable approach in this situation $\endgroup$ Commented Apr 25 at 14:09
  • 2
    $\begingroup$ Please don’t delete and then re-ask the same question. This destroys the work that people have contributed to improve the question. Instead, edit the question. $\endgroup$ Commented Apr 25 at 14:59

0

Reset to default

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.