overlapping in PE library sequencing

[hi-koike]

Hello,

I'd like to do de novo assembling of a fungal genome(~20 Mbp).
I will use Miseq data for the assembling.

I found some programs can handle overlapping sequence of the two
paired end reads such as ALLPATHS-LG and FLASH.

I am wondering how long of the fragment length for paired read is appropriate for the version2 of Miseq.
Now the read length of the Miseq v2 is 250 bp, paired reads from fragments of size ~400 bp or more would be overlapping. On the other hand, the bases in the terminus might be low quality and can not be used, right?

I guess longer fragments would be effective in getting information from the same number of reads, while it would be meaningless if the reads would be too short to overlap due to their low quality.
Has anyone try to input overlapping fragment library from the fragments of
more than 400 bp for any de novo assembling ?

Many thanks,

[JamieHeather]

I recently had my first play with combining overlapping reads (as detailed in this thread).

I was trying to overlap amplicons of one of two lengths, sequenced on a MiSeq v2 2x251 kit. Bottom line, the 420bp fragments overlapped great, the 320bp fragments wouldn't combine at all (unless I trimmed the reads down a bit so that one read wasn't extending past the start point of the other).

So at the very least, an overlap ~70bp is plenty to get > 90 % of reads overlapping (note that the mean quality at the end of my reads is still > Q20, just).

[hi-koike]

Hello Jamie,

Thank you for your reply. The thread is helpful for my understanding.
Your blog is also helpful and interesting to read.

Thanks,
Hideaki