Compare two scatter plots? Statistical test?

Question

[Ankit]

Hi everyone,

I have two scatter plots (please see figure ), one diseased other rescued condition. I wanted to show that methylation loss which has occurred in diseased (red line plot left) is recovered to some extent upon rescue (blue line plot right).

Any suggestions which statistical test need to be applied to show that overall rescue has occurred?

The rescued cells were derived from diseased cells.

Thank you

[Shraddha]

I'm not sure about the relevance of the visualizations you've provided to the statistical tests. The simplest thing you can do is run a student's t-test on the datasets you used to generate these.

[Ankit]

Please elaborate how. I believe ATpoint has a good suggestion for fitting model. It has something to do with linear regression.

[Mensur Dlakic]

You have shown two plots with two lines in them. Scatter plots have a scattering of data points in them - hence the name.

[ATpoint]

Is this NGS? The regions are the same in both conditions? Maybe some kind of goodness-of-fit test. Please add details on the experiment.

[Ankit]

It is an array data. Plotting is done using average methylation of probes under 500 bp windows. Methylation value ranges from 0 - 1

Regions are same in both condition for solid and dashed lines. I tried to fit a model using lm function in R. But was not sure how to use the output and compare the two.

Do I need to fit a model for dashed and solid lines separately and then compare output of lm like Coefficient? Residual? P-val? among dashed and solid lines between two scatter?

Or

Should I compare lm output within the scatter for dashed and solid lines and take some value from each scatter and compare. Please suggest.

Thank you

[Ankit]

In R, I applied

model <- lm(y ~x + condition, data = data)

summary(model)

Please let me know if it is ok?