This shows you the differences between two versions of the page.
Both sides previous revision Previous revision | |||
lecture_notes:04-19-2010 [2010/04/24 08:08] cbrumbau Merged rest of notes, expanded some acronyms |
lecture_notes:04-19-2010 [2010/04/24 14:54] (current) karplus added correction about quadratic behavior of overlap-consensus |
||
---|---|---|---|
Line 34: | Line 34: | ||
* Have to compare all reads against each other. | * Have to compare all reads against each other. | ||
* Quadratic. This has become difficult. | * Quadratic. This has become difficult. | ||
+ | * Correction: quadratic behavior can be reduced to about O(n log n) by doing a clustering before aligning: just like in several fast search techniques, you only attempt to align sequences that share seed hits to relatively short k-mers (much shorter than those needed for de Bruijn graphs, since they only need to reduce the number of hits, not get unique locations in the genome). --- //[[karplus@soe.ucsc.edu|Kevin Karplus]] 2010/04/24 07:51// | ||
* Next generation sequencing reads are short, and overlaps between reads are short. They are getting a little longer with newer technology. | * Next generation sequencing reads are short, and overlaps between reads are short. They are getting a little longer with newer technology. | ||
* OLC method uses all information in the read at least. | * OLC method uses all information in the read at least. |