Noise Removal without affecting peaks

I acquired voltage waveform with noise. I tried different filters to remove noise but it also affects peak levels. Please help me in this regard. I have attached my data file too.

[John D'Errico]

Why would it not impact the peaks? They are as likely corrupted by noise as is any other bump in that curve.

[Mubashir]

but filters have too much impact on peak amplitude because of high noise. Peaks are the key features in my algo.

[John D'Errico]

But ANY type of smoothing/filtering/noise removal process will treat high peaks that are uncharacteristic of the rest of your curve as potentially noise. So any such method will always try to reduce those peaks. And surely if your data is as noisy as it appears, then there is no reason to believe that a peak is not just as likely to have noise in it too.

If you cannot tolerate something playing with those peaks, then you cannot tolerate any noise removal scheme.

[Mubashir]

Thanks for response. I am still trying to sort something good for given data.