Sequencing strategey for large fish genomes and AllPaths assembly

[spavey]

We are starting a number of large (1.5-4 Gb) fish de novo genome sequencing projects. The AllPaths-LG assembler seems to be producing the best genomes of these types. I am wondering what is the best stategy for overlapping PE libraries for an AllPaths assembly. We were going to sequence 6 lanes of Hi-Seq rapid mode, to get around 100x coverage. Is it best to make one PE library and spread it over the 6 lanes, one library for each lane, or something in between?

Thanks

[dfhdfh]

In general, I made the experience that spreading individual samples over all lanes is quite beneficial. In one run, we had a defective lane and still got our (reduced) data because we multiplexed all samples across all lanes. These, however, were RNA-seq data.

[jfostel]

Is six lanes of sequencing the amount you will need total for this experiment, i.e. you are only sequencing one individual and 6 lanes will yield the total coverage you need for assembly?