stranded or non-stranded protocol?

[enelkinsan]

Hi, all
we've got a differential expression study on some poorly annotated organism (fish), and the task is to compare expression in different tissues. I was wondering, may be someone has any recommendations on the type of protocol to use? We are free to choose between stranded or non-stranded, but I don't really know the pro and contra of each...
Or point me in the direction of right paper, if you got one in mind.

Many thanks!

Elizabeth

[cmbetts]

My uninformed opinion

My understanding is that a stranded protocol is better for poorly annotated organisms because the extra strand information helps to more accurately assign the reads when reconstructing the transcriptome (Although I have not personally done this myself.
Also, unless there were other considerations involved, it seems that stranded gives you more information for you sequencing buck.

[enelkinsan]

Thanks for that!
By the way, it makes me thinking - why would people need non-stranded protocol than? There must be some "pro-s", or it would just die off...

[Dario1984]

A stranded protocol is more complex, and therefore more expensive to buy. But, new versions of sequencing kits are stranded. I don't think unstranded kits will be sold much longer.

[enelkinsan]

Ok, on today's meeting there came a reply to my question: apparently, non-stranded protocol is preferable if transcripts are likely to appear on both strands.
In all other cases stranded goes best.
It was also said that in terms of money to spend they do not differ much.

[pmiguel]

Ok, on today's meeting there came a reply to my question: apparently, non-stranded protocol is preferable if transcripts are likely to appear on both strands.

I don't get it. If you have transcripts on both strands, then a a stranded protocol will collect that information for you. Everything else being equal, why would you not want that information?

The reason not to do stranded libraries would be if they were too expensive or otherwise difficult to produce. Otherwise the strand information will either be useful or can be ignored.

We have tried the new Illumina Stranded kit. Seems just as easy to use. Probably we will discontinue use of the non-stranded kit.

--
Phillip

[pbluescript]

Ok, on today's meeting there came a reply to my question: apparently, non-stranded protocol is preferable if transcripts are likely to appear on both strands.
In all other cases stranded goes best.
It was also said that in terms of money to spend they do not differ much.

Like pmiguel said, that comment doesn't make sense.
Where stranded hasn't caught up is in dealing with small amounts of RNA or highly fragmented samples. Some of my samples simply can't be done with the stranded protocols out there.

[enelkinsan]

Well, I didn't get it either, I just translated someone's remark, as he was quoting someone else. If I ever run into that quoted guy, I'd ask him. Meanwhile, we are turning to stranded protocol anyway