Hello all!

Case:

We did a Sequencing of three yeast strains (not s.cerevisiae) and with reference genome available!

Illumnia MiSeq

around 8,5 million reads each way (paired ends)

most reads around 90-120 bp long

high coverage: around 90x

In the trimming /adapter removal step we saw most of our paired end reads have overlaps

R1 ------------------------->
R2 <-----------------------

99% overlapping paired end reads 1% non-overlapping paired end reads

Question:

Is it better for the next assembly steps (initial contig building, scaffolding)

-> to only use single-end reads: Collapse the overlapping paired end reads (99 %) into single end reads (since assemblers can have problems with overlapping paired end reads) and use only this single end reads for the assembly (discard the 1%) ?

-> to only use paired-end reads: use the overlapping paired end reads (99%) and the non-overlapping paired end reads (1%)?

-> to use a mix: single and paired-end reads: Collapsed into single end reads (99%) and non-overlapping paired end reads (1%).

Thanks