how to order the reads of DNA methylation data from bisulfite sequencing

[Jim Michael]

How do we order the reads after sequencing ?
1. Can we label them by numbers?
2. Can we label them by the real-time?

For example, we consider: reads from sequencing sorted by mean methylation.
But randomness of reads leads to different DNA methylation patterns, thus different classes. It could also cause other research errors.

Thank you very much for your suggestions!

[chadn737]

This doesn't make sense. Are the reads aligned? What is the purpose of "ordering" them?

[Jim Michael]

Thanks for your questions!
I just feel this question is very interesting.. :-)
Yes, they are homologous, after alignment we can get the consensus sequence..and also the mean methylation of this sequence..
without ordering it is difficult to analyse DNA methylation profile of this specific gene..

[chadn737]

What do you mean "ordering" and why does that affect you're analysis.

[Jim Michael]

We might have hundreds or thousands of reads for one amplicon, thus after alignment with the reference gene, the positions of CpG site are randomly distributed in all of the reads. In order to study DNA methylation pattern (different positions of CpGs in different reads) we need to order them first..
ps: here we consider only CpGs sites..

[chadn737]

If you already have done an alignment then you know the position of each CpG relative to the reference gene. For methylation then you are interested in the number of methylated and unmethylated bps covering each CpG in you reference.

[Jim Michael]

you are right, this is part of my interest.
Perhaps you understand DNA methylation pattern (like, 2D graph) differently.
Let say, for one amplicon, we have sorted reads in sample A and sample B, but random in sample C and sample D, if we want to do clustering of samples, how can one draw a reasonable result?
cheers :-)

[chadn737]

Don't you want to cluster based on methylation levels? Your reads shouldn't have to be sorted for this.