Hi, I have a large signal (coverage file) as below and my goal is to

1. detect the change points in the coverages
2. identify change points that have gradual increase and decrease in coverages (point C1 and C8)
3. identify points where there is a substantial increase in coverages (C6 counts, while the dip between C3-C4 does not count)

May I know what are some good change-point detection algorithm/ deep learning methods that would allow me to achieve all these goals accurately?

Any input would be incredibly appreciated as I am very new to this field, thank you!

• Any suggestions on how to normalize/smooth my data would be very helpful too - So far I am normalizing it by dividing all the points by the overall median, and am applying the the Savitzky-Golay filter to smooth it.

 

Your requirements are probably unique. I doubt whether anything exists to do exactly what you want. You should start by defining all of your requirements in detail. From that definition you should be able write a logical algorithm to achieve your desired results.