Effect of wide peak width on RNASeq

[Noa]

Hi
I am working with an NEB kit to prepare Illumina samples. After some troubleshooting with their reps, they recommended an alternate protocol whereby instead of two size selection steps after adapter ligation, I just do two rounds of AMPure bead cleanup.
Bottom line of this is that instead of getting a library from 200-400bp with a beak at ~240bp after the final PCR, I get a smear with roughly the same peak but the width is 200-700bp (see attached figure).
Any thoughts on how this may affect downstream bioinformatics? I need to do two things with the libraries - one is a differential expression analysis between samples; the other is transcriptome assembly and then DE (two completely different experiments)
thanks!

[rebyl]

I would also like to know the answer to your question! I have a broad (~200-900bp) distribution of fragment sizes (see Bioanalyzer traces attached) and am concerned about how this will affect assembly of my reads. I will be doing do novo assembly using Trinity.

[Noa]

Hi rebyl-
I had done my trinity build on peaks that were smears from 200-700 in the end. I had written the trinity rnaseq users group and brian haas answered me as copy-pasted below. The trinity build from these data worked beautifully in the end (using default parameters). FYI Brian is super-useful knowledge source and answers daily! So use the trinity rna seq email group if you run into more issues that are trinity-related!
noa

his email:
I expect it should be fine with the default parameters. If you'd like, you can set the --group_pairs_distance <int> value higher than the default (500), which could help for a small number of transcripts, but will unlikely impact the bulk of the assembly.

best,

-brian