What is indexing a genome?

[bingalee]

Hello folks,
This question might come across as naive, but I'm working on Rna SEQ data, and I want to get my basics right.
I used the tool bowtie for indexing my reference genome and then used TOPHAT to aling my rna-seq data to the reference genome.
I dont understand what exactly, "indexing" of the reference genome is and why it had to be done.
Can someone please explain it to me?
Thanks in advance

[BAMseek]

Aligning reads to a genome can be viewed in the general context of approximate string matching. The goal is to find a pattern (the short-read) in a large text or corpus (the genome), allowing for mismatches and indels. Naively, you can scan the text for the pattern but this is inefficient (think of trying to find an address in a phone book where the names were all mixed up). There are techniques to pre-process (or index) the text to make queries fast and also that can even compress the size of the text. At the heart of Bowtie is the Burrows-Wheeler transform, which is used to compress the size of the genome and also make searching the genome faster.

This thread gets into more details.

Justin

[GenoMax]

See this post for supplemental information.

http://blog.avadis-ngs.com/2012/04/elegant-exact-string-match-using-bwt-2/

[maivantan]

how many time we should do index reference genome? I mean that can we index one time for all alignment or each alignment we have to index reference again?
Thank you

[Brian Bushnell]

how many time we should do index reference genome? I mean that can we index one time for all alignment or each alignment we have to index reference again?
Thank you

You only need to index a genome once, until a new version is released.