This is a test version of Biostars. For the public version, visit https://www.biostars.org.
Choosing right statistical test in genomics

I have a bed files contain coordinates of one specific type of repetitive DNA in a chromosome. e.g.

0   190 TEsI
2844    2881    TEsI
2881    2988    TEsI
...

Now I'm interested in few specific regions, they about 10 kb in length. My supervisor asked me to comment on whether or not there is statistical significance in number of these types of repetitive DNA in these regions. I've plotted this: enter image description here

Red stripe shows the position of region, horizontal black and green dashed lines indicate mean value and mean value with standard deviation for the whole chromosome with bin sizes of 100 kb (because this plot covers about 100 kb). How to test null hypothesis (there is no significance) here? Thank you in advance

repeats genomics annotation statistics bed
• 1,537 views • link
updated by Jean-Karim Heriche Jean-Karim Heriche • written by kirillkirilenko • kirillkirilenko

2 answers

You can use a one-sample t-test or a one-sample z-test. Both these tests determine whether there is a significant difference between the mean of a sample (number of repetitive DNA in regions of your interest) and a known or hypothesized population mean (mean value of repetitive DNA in the whole chromosome).

Null hypothesis: the mean of the repetitive DNA is 1 (the value of the horizontal black line).

Alternative hypothesis: the mean of the repetitive DNA is not 1 (the value of the horizontal black line).

• 0 views • link
written by Andrzej Zielezinski Andrzej Zielezinski

Isn't the difference in bar size sufficient for you to declare that this is relevant to the biology you're interested in? This question isn't going to be answered by a statistical test. It seems you're looking at counts in intervals so I believe you should be looking at discrete statistics like based on the Poisson distribution especially because counts seem to be low.

• 0 views • link
written by Jean-Karim Heriche Jean-Karim Heriche

Log in to answer this question.

More posts like this