de novo sequencing of maize (lane and insert size library)

[sflong]

we want to sequece a self-cross line of maize, of which repeat sequnces accounts about 70%~80%. we plan to construct mult-insert size libraries, including 300bp, 500bp, 800bp,2000bp,5000bpand 8000bp and utilize 4 lanes to deliver about 50X data using solexa platform. so my question is
1) how to assign different insert size libraries to the 4 lanes?
2) does data of mate pair libraries produce a lot of redudant reads?

thanks a lot!

[sflong]

our plan is 300bp--- 1 lane,
500bp--- 1 lane,
800bp--- 1 lane,
2k,5k,8k mixed---1 lane,
I wonder if anybody have better idea for the project?

[pmiguel]

I have been wanting to try a DSN normalization method for a genome like maize. Nearly 50% of the maize genome comprises a handful of LTR retrotransposon families. Seems like you could wipe them out with a few cycles of denaturation-annealing followed by DSN.

Then you get a de novo assembly without most of the retrotransposons (which are not going to assemble well anyway using full genome shotgun).

--
Phillip