RIN of 6- should I run sample?

[Noa]

I have RNA from hard-to-come-by samples. Some RINs are 8+ but some are 6.3-6.9. I would need comparative RNA-Seq work on this, as well as transcriptome assembly, as this is an unsequenced genome.
Should I chance the samples with the low RIN on an Illumina?

[pbluescript]

Go ahead. The first step in many RNA library prep protocols is fragmenting the RNA, so a lower RIN is generally not a problem. RINs in the 6 range really aren't that low.

[mnkyboy]

If it is whole transcriptome it will not matter, if you are doing some mRNA-seq you may see some loss but if it is for assembly you should be fine since you are probably doing a lot of reads.

[pmiguel]

RINs that low will cause a problem in most use cases.

(1) If you plan to do a polyA+ RNA purification (as in TruSeq RNA prep libraries), then only the segments of your RNA still attached to their polyA tails will be isolated. The result will be 3' bias in your sequence reads.

(2) If you plan to ribo-deplete your sample using a hybridization-based method, then some ribosomal RNA will "bleed" through the process because the ribosomal RNA will be somewhat fragmented. Hybridization-based methods generally rely on a set of biotinylated oligonucleotides complementary to certain regions on the ribosomal RNA. Extensive fragmentation will create fragments without these binding sites on them and allow them to escape the depletion process. The result will be a higher level of ribosomal RNA contamination of your reads. The result will be fewer useful reads. Epicentre's "ribozero" is more effective at removing slightly degraded ribosomal RNA -- presumably by supplying a more generous diversity of binding oligos.

(3) If you plan to use normalization to ribo-deplete your samples, then everything is likely fine. Except that normalization is trickier than other methods of ribo-depletion.

--
Phillip