1
$\begingroup$

I have time series data from NanoPore sequencing (Attached is a illustrative figure and short explanation) which I'd ultimately like to use in order to find various methylation patterns in the input DNA sequences.

For those of you for whom the above paragraph doesn't mean much: This is a hidden Markov model sort of situation, where the data is some noisy function of the true input sequence, and I would like to infer certain properties of said input sequence.

I have a few approaches in mind. But first, what would be some some good data exploration methods? Especially ways to visualize the raw data that would help in understanding the type of structures which are present. For example, I thought of dividing the data into overlapping windows and plotting a matrix of [Pearson?] correlation between the different windows.

Any suggestions and/or references for further exploration are greatly appreciated.

NanoPore sequencing: This is a DNA sequencing method in which the DNA strand is forced through a "hole" in a membrane while at the same time there is a flux of Ions from one side of the membrane to the other. Different nucleotide patterns block the hole in different ways, creating a measurable difference in the flux, which is being measured constantly (vicariously, via the current-density) on the membrane.

$\endgroup$
2
  • $\begingroup$ your data is interesting . Are there predictor variables that you can suggest ? Solutions to time series problems are often based on actual data as theory can be often too general to be useful. $\endgroup$
    – IrishStat
    Apr 30, 2019 at 8:06
  • $\begingroup$ this is often called EDA ...Exploratory Data Analysis $\endgroup$
    – IrishStat
    Apr 30, 2019 at 8:12

1 Answer 1

0
$\begingroup$

If you are interested in some basic tools to visualise and explore the data, I published a toolkit in 2019 called SquiggleKit to do that.

https://github.com/Psy-Fer/SquiggleKit

For methylation analysis, you might be interested in nanopolish evenetalign or tombo resquiggle.

Alternatively, there is now a new secondary network used to decode methylation states during basecalling called remora https://github.com/nanoporetech/remora

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.