Sequencing data analysis of pair-end sequencing

[Smriti]

Can anyone suggest me an article on DNA sequencing data which can explain me terms like: Aligned Pairs (Proper), Aligned Pairs (Improper), Aligned singletons, Aligned sequenced. It would be a gr8 help for beginner like me.

[ECO]

Do you understand what's going on in paired end sequencing? I know that's not your question but those categories above should be nearly self explanatory.

I'm not sure what context those exact terms are from, but the "proper"/"improper" refers to the relative orientation and distance between two paired end reads.

Proper identifies a pair of reads maps to the reference at a distance within the expected distribution (defined by the library construction) AND in the correct orientation to one another on the reference genome (defined by both the library construction method and the sequencing technology).

Improper would mean that the pair violated one or both of these specifications...perhaps meaning there is a structural variation at the location of interest.

Singletons would just be one of the two members of a pair was not aligned...could be a consequence of a low quality read or some other consequence of the data collection.

Couple other good threads to read...good luck!

Basics: http://seqanswers.com/forums/showthread.php?t=503
Orientation: http://seqanswers.com/forums/showthread.php?t=3214

[Smriti]

Thanks ECO for all this help!

I understood this to some extent, if not fully. I have another beginner's doubt. How do you describe "our" genome and "reference" genome when we compare. If its genome of same individual then how are deletions and insertions introduced in "our" genome when we map it with reference genome. Should not it be exactly the same. Or reference genome is same species but different individual. Can you please explain this.

Thanks again!

[gringer]

The reference genome is whatever genome you're mapping onto. In the case of human, it might be NCBI build 37.3, or UCSC hg19 (I think they're essentially the same), or an assembly from a different group. Every reference genome will have its own set of variants that it considers to be the most likely to be "normal", and those are what you map to.

[Smriti]

Thanks gringer,

Apart from human genome, does that mean every species has some reference std. which is considered as reference genome?

[sophied]

I have found that this paper is quite useful for people starting off in this area:

http://www.ncbi.nlm.nih.gov/pubmed/21749903

[gringer]

Apart from human genome, does that mean every species has some reference std. which is considered as reference genome?

If you're mapping to a reference genome, then you must have a reference genome that you are using for the mapping. This will not necessarily be the same as the reference genome used by someone else.

[Smriti]

Thanks gringer and thanks sophied!

I have downloaded the article and would be reading it now. Please suggest me more articles where i can get good picture of pair end sequencing and data analysis specially working with human genome.

[westerman]

Thanks gringer,

Apart from human genome, does that mean every species has some reference std. which is considered as reference genome?

I am rolling around the floor in laughter.

One of my biggest headaches is trying to find good reference sequences for all of the non-model organisms we sequence. It isn't easy. Even for the reference organisms there are often multiple ones -- see http://cufflinks.cbcb.umd.edu/igenomes.html for a list. Let me know which of the cow ones you would like me to use for mapping versus sheep (because there is no good reference for sheep).