how to use Butterworth filter to get rid of noise?

I cannot figure out how to use a lowpass butterworth filter on this piece of data.

Here is what I have so far

Ts = mean(diff(x));
Fs = 1/Ts;
Fn = Fs/2;

I found the sample rate to find Wn, however I do not know what to set the cutoff value.

Wn1=10/(Fn);
[b,a] = butter(6,Wn1,'low');
y2 = filter(b,a,y_pr);

unfortunately this would distort the data and make most peaks the same height

[Star Strider]

Ahmed Emam later added:

‘... my code is also required to handle a sampling rate varying from 0.5 kHz to 2 kHz. I wanted to use the lowpass option in the butterworth filter. The output I am looking for is the same EKG graph but with less noise.’

This is not a trivial omission.