Multiplexing without indexing sequencing

[spenserbsmith]

Dear board,

I am trying to develop a method for adding 6mer barcodes to RNA-seq protocols with the illumina platform. From what I understand, indexing under illumina's methodology has an entire adapter and reaction to read the 12 different sequences that illumina chose for their specific indexing. Is there an easy way to insert my own hexamer barcode which would be read under the main sequencing reaction i.e. in-between the bridge PCR adaptors and the library strand. We would be doing paired end sequencing and I would be using a few different protocols initially to test efficacy (while our lab chooses which method for our RNA seq to use) and need an easy method to simply insert our desired barcode.

Does this make sense? Any help?

I have initially thought that we could do a site directed mutagenesis insertion but I don't know how cheap or practical that is for our RNAseq.

[spenserbsmith]

Such as this:

http://www.clontech.com/products/detail.asp?tabno=2&product_id=10586

[spenserbsmith]

This is much easier to just change all the adapters and include a 3mer barcode for multiplexing. As well you must include the extra T in your pcr amplification primers and adaptors so the sequencing adapters still work.

I'm still trying to figure out how to do this on ScriptSeq. If anyone has ideas, feel free to post!

[upendra_35]

This is much easier to just change all the adapters and include a 3mer barcode for multiplexing. As well you must include the extra T in your pcr amplification primers and adaptors so the sequencing adapters still work.

I'm still trying to figure out how to do this on ScriptSeq. If anyone has ideas, feel free to post!

Hi i am also in the process of figuring out how to multiplex the 3mer barcodes on ScriptSeq in similar lines to Illumina which we have done successfully. One way we can do it in SS is to use 3-mer barcode along with 6N primers and use as RT primer. What do you think about this?????

[scooter]

I think the reasons for placing the barcodes in a distinct location away from the main sequencing read is that it results in far less representation bias by adding the barcode by PCR instead of by ligation or first strand cDNA synthesis.

If you try and modify ScriptSeq by adding a barcode onto the randomer you will risk having selective priming events at the RT step that might throw off gene expression measurements. Likewise, if you add barcode to small RNA adaptors and use them for ligation you will likely have ligation bias prefer certain barcodes more than others and again throw off representation of the transcripts in the sample.

[upendra_35]

I think the reasons for placing the barcodes in a distinct location away from the main sequencing read is that it results in far less representation bias by adding the barcode by PCR instead of by ligation or first strand cDNA synthesis.

If you try and modify ScriptSeq by adding a barcode onto the randomer you will risk having selective priming events at the RT step that might throw off gene expression measurements. Likewise, if you add barcode to small RNA adaptors and use them for ligation you will likely have ligation bias prefer certain barcodes more than others and again throw off representation of the transcripts in the sample.

Thanks for the useful info. I was in the impression that the priming will be random because of '6N' nucleotides attached to the barcode. If you think it might still cause selective priming events, do you have any other option for multiplexing. At the moment we can't multiplex the way Scriptseq advises in the protocol because our Sequencing facility does not use MIR primer to sequence the barcode and the only way we can multiplex is to use the barcodes at RT step. Advise me if you have any thought on this.... We definitely want to try Scriptseq kit to compare it with Illumina.......

[scooter]

Yea, one would assume that the N's are the major priming substrate, however the remaining sequencing in the first strand primer will certainly influence the sequence context captured. If you use only one primer sequence than every sample will have the same bias. However, add a 6 nt barcode and you now have a good chance of altering that bias since the barcode could contribute to the selection of priming site, thus resulting in a misrepresentation of read counts. I do not know for sure whether this will happen, I am only pointing out that it is a logical risk that should be considered.

My advice is that if you want to use the barcodes as you describe I would do a small scale pilot study and barcode the same exact RNA sample with whatever number of barcodes you want to use. Analyze the concordance of each barcode with the others and see if any particular sequences skew the results. If the correlations between barcodes are very, very high and there is little to no difference, than you are in good shape. This might also tell you that one or more of the barcodes should be dropped because they behave strangely.