[sidderb]

Does anyone worry about the quantification of expression from total RNA versus ploy A enriched RNA?

I imagine that you would inflate your expression values from total RNA by including reads from immature RNA and I presume that as splicing is a regulated event this might not be even across genes. But, you would not discriminate between immature and mature RNA when doing qPCR from total RNA (the gold standard), so does this matter?

[NextGenSeq]

The reason people get more data mapping to transcripts with methods using oligo-dT reverse transcriptions is that you ONLY see polyadenylated transcripts i.e. the boring ones. You might as well do a microarray. Oligo-dT selection or priming gives horribly 3' biased data and I never use it.

The only way to make an RNA-Seq library is through ribosomal depletion and random priming or the SPIA selection method. This gives data which is not 3' biased and you see the non polyadenylated transcripts AND the alternatively spliced transcripts. With 3' biased libraries you miss both.

[Jon_Keats]

The reason people get more data mapping to transcripts with methods using oligo-dT reverse transcriptions is that you ONLY see polyadenylated transcripts i.e. the boring ones. You might as well do a microarray. Oligo-dT selection or priming gives horribly 3' biased data and I never use it.

The only way to make an RNA-Seq library is through ribosomal depletion and random priming or the SPIA selection method. This gives data which is not 3' biased and you see the non polyadenylated transcripts AND the alternatively spliced transcripts. With 3' biased libraries you miss both.

Can't say we see any significant 3' bias in our TruSeq mRNA-seq preps (poly-A selection followed by heat fragmentation and random priming). If anything this is more common in our Nugene preps.

[pmiguel]

Yes, I think NextGenSeq's comment is an exaggeration. With high quality RNA, 5'-3' bias is minimal when using the TruSeq RNA kit. Oligo dT priming of first strand synthesis is another story; it would likely produce bias because reverse transcriptase is generally not very processive.

On the other hand, if polyadenylated messages are "boring" to you, you must use another ribo-depletion method.

--
Phillip

[NextGenSeq]

I've seen UCSC browser plots of RNA-Seq data which map virtually entirely to the last exon using dT priming.

[NuGEN]

Input of total RNA vs poly(A)+ RNA

After reviewing the various posts in this thread, I wanted to clarify some points regarding the function of Ovation RNA-Seq System V2. In my view, the choose of input is guided by the experimental goals of the RNA-Seq study. The Ovation RNA-Seq System V2 was designed for use
with total RNA (500 pg minimum input); however, some studies may warrant use of poly(A)+ RNA. This is particularly true for screening
or profiling studies examining expression under different
biological conditions using a defined set of polyadenylated
transcripts. Researchers particularly interested in RefSeq
or other well-annotated transcripts or who want to enrich
for sequencing reads from mature transcripts may also find
it desirable to use poly(A)+ as input.

We have recently produced an Application Note (http://www.nugeninc.com/nugen/index....brary-systems/) showing how poly(A)+ RNA can be used as an alternative to total RNA if the experiment calls for use of this material. This note shows data both for Ovatin RNA-Seq V2 as well as Encore Complete RNA-Seq Library System. This latter kit also provides strand tracking data which may be useful for those interested in antisense trancripts.

Steve

[stighe]

Mapping to reads in non-coding region is a function of poor DNase treatment. My guess is the user did not check their RNA for genomic DNA by qPCR. A obvious mistake.

[mulligan]

I've seen UCSC browser plots of RNA-Seq data which map virtually entirely to the last exon using dT priming.

Sounds like the RNA from that experiment was degraded or maybe it was DGE?

[geneart]

multiplexing strategies for GA/Illumina

Hello ,
I am just getting warmed up with NGS and came to understand that while multiplexing samples on GA, we need to watch out for the indices that we pool together. I was thinking so far, that as long as the indices are different we can pool as many as we can per lane. So this new thing threw me off . Also I read this little snippet , the link is posted, which gives the adaptor index pool compatibilites. NOw, all that I have is adpator for 3' and 5' along side my RNA which I used to perform an RT and PCR. While doing the PCR I added the indices 1-12 according my protocol . I have 48 samples which all have unique indices. When I try to pool I would make sure no indices overlap.
SO can someone clarify what the guys are talking about in the link I have posted below? Or am I totally on a different planet than earth!!!!!!!
Thanks,
geneart.TruSeq_SamplePrep_Pooling_ProductInsert_15034135_A.pdf
TruSeq_SamplePrep_Pooling_ProductInsert_15034135_A.pdf

[geneart]

adding to my previous post................Also, I cross checked the "AD" index numbers listed in the attachment I posted in my previous blog,against illumina truseq protocol and could not find a match for the catalog #.
Now I believe there should be something I am totally overlooking !!!! Atleast until the part of RTPCR I am 100% sure that other than the adaptors for 3' and 5' and the indices(which have a catlog number that does not match AD numbers) I have not added anything extra !!!! So where do these AD006 and other AD numbers come from?? Is it something that we add downstream before running GA? According to my understanding , the next few steps invloves concentration check and getting eqvimolar concentrations ofor all samples and running the GA!
Also the indices itself is the "barcode" to identify the tags, Is'nt it?
Please can anyone help me get this right?
Much appreciated !!!!
Thanks
geneart

[pmiguel]

Yes those are "Low plex" instructions. That is, in cases where you want to use 4 or fewer indexes in a lane. Above that number you will be fine, most likely.

Also, this seems to be an issue for the GA-IIx, the HiScanSQ and possibly the MiSeq. Near as I can tell the HiSeq has no difficulty with low plex pools. Ours demultiplexed a 1-plex pool at >98% efficiency.

The above comments concern TruSeq indexes. Nextera indexes may be another issue, but I doubt it.
--
Phillip

[geneart]

THanks Phillip !
But then what are those AD and AR numbers? Do they correspond to the adaptors? THat does not make any sense
Also I have a scenario where I might multiplex (>5) per lane in few lanes and run just one(uniplex) in one lane. So just because it is uniplex and not with any other combo it should be fine????
I would appreciate if you could throw some insight in dealing with samples with <50ng/ml(qubit) and 80pg/L on agilent. The molar concentrations need to be 10nM for NGS/GA run, we are trying to come up with strategies to make these runs even with low concentrations by drying them and such. Even so I am afraid we may not have enough. So I thought we could pool strong samples in some lanes, again checking the indices compatibility and just run one low conc. sample in one lane alone. Would you have any experience dealing with very low concentration samples on GA?
Much appreciated...............
geneart.

[pmiguel]

THanks Phillip !
But then what are those AD and AR numbers? Do they correspond to the adaptors? THat does not make any sense

The "AD" prefix means the adapter came from a "DNA" prep kit, whereas "AR" means they came from an RNA prep kit. Your sequencer has no reason to care about this, because the same adapters are used in both types of TruSeq kits.

Also I have a scenario where I might multiplex (>5) per lane in few lanes and run just one(uniplex) in one lane. So just because it is uniplex and not with any other combo it should be fine????

Yes. Well, if you spike phiX in the lane, that will show up, unless you demultiplex.

I would appreciate if you could throw some insight in dealing with samples with <50ng/ml(qubit) and 80pg/L on agilent. The molar concentrations need to be 10nM for NGS/GA run, we are trying to come up with strategies to make these runs even with low concentrations by drying them and such. Even so I am afraid we may not have enough. So I thought we could pool strong samples in some lanes, again checking the indices compatibility and just run one low conc. sample in one lane alone. Would you have any experience dealing with very low concentration samples on GA?
Much appreciated...............
geneart.

Sorry, we never had a GA. We do speed vac samples sometimes to concentrate them a little. But we only need 2nM for the HiSeq and MiSeq.

--
Phillip

[geneart]

NGS multiplex

Thanks so much Phillip ! It was very helpful
geneart