HiSeq multiplexing for strand specific RNA library prep
Greetings, fellow scientists!
I am a grad student and my first task in lab is to pioneer a gene expression experiment to be run on the HiSeqSQ. I will be using RNA from mouse brain, I want to multiplex, and I'd like to obtain paired-end reads. I have lots of questions and any help would be greatly appreciated!
I'm intrigued by a recent protocol published by CSH, 'High-Throughput Illumina Strand-Specific RNA Sequencing Library Preparation' by Zhong S et al (http://cshprotocols.cshlp.org/conten....prot5652.long), which reduces the cost of library prep to $10/library - much cheaper than the TruSeq kit!
The authors explain that using the Craig indexing method (2008 - Nature Methods) is not suited to the HiSeq platform because the dominant 'T' signal from the Craig adapter T/A overhang interferes with its base calling/focusing system if it appears in the first 5 cycles (apparently this was not problematic for the GAII). Thus, they modified the length of their barcodes, to prevent having only T in a particular cycle. They used a mix of 4 and 5 nucleotide barcodes in their experiment.
In a recent paper by Kircher et al (2011) - BMC Genomics, I read that the Illumina base caller uses two parameters (phasing/prephasing value and cross-talk matrix) to convert intensity values into bases. These estimates are often inaccurate when base composition is unbalanced. But Kicher et al claim that this method is not used for index reads: "for the index, parameters calculated for the preceding read are applied. Therefore for at least one lane in each run the base composition should be balanced over the thousands of clusters in the tile OR a separate control lane must be sequenced for estimating these base calling parameters." Does this mean that if I am running multiplexed paired-end samples on a HiSeq machine, I can use only 6-nt Craig adapters if I also include a Phix control library? Or is the problem still the same?
I'm confused about this. Basically I am trying to figure out if I can side-step the process of designing my own 4-5nt barcodes since the 6nt ones are already available in the Craig paper. Does anyone know of any references which include sequences for shorter (or longer) barcodes that I could order from IDT, or have tips on how to design them?
I found a thread on multiplexing with Illumina adapters/primers which was very helpful (relevant document attached), as it listed the sequences for Illumina adapters, multiplexing primers, etc. Which ones would be appropriate for paired-end multiplexing? I assume the multiplex ones listed in the document are for single-end reads.
Thanks!
Natalia
I am a grad student and my first task in lab is to pioneer a gene expression experiment to be run on the HiSeqSQ. I will be using RNA from mouse brain, I want to multiplex, and I'd like to obtain paired-end reads. I have lots of questions and any help would be greatly appreciated!
I'm intrigued by a recent protocol published by CSH, 'High-Throughput Illumina Strand-Specific RNA Sequencing Library Preparation' by Zhong S et al (http://cshprotocols.cshlp.org/conten....prot5652.long), which reduces the cost of library prep to $10/library - much cheaper than the TruSeq kit!
The authors explain that using the Craig indexing method (2008 - Nature Methods) is not suited to the HiSeq platform because the dominant 'T' signal from the Craig adapter T/A overhang interferes with its base calling/focusing system if it appears in the first 5 cycles (apparently this was not problematic for the GAII). Thus, they modified the length of their barcodes, to prevent having only T in a particular cycle. They used a mix of 4 and 5 nucleotide barcodes in their experiment.
In a recent paper by Kircher et al (2011) - BMC Genomics, I read that the Illumina base caller uses two parameters (phasing/prephasing value and cross-talk matrix) to convert intensity values into bases. These estimates are often inaccurate when base composition is unbalanced. But Kicher et al claim that this method is not used for index reads: "for the index, parameters calculated for the preceding read are applied. Therefore for at least one lane in each run the base composition should be balanced over the thousands of clusters in the tile OR a separate control lane must be sequenced for estimating these base calling parameters." Does this mean that if I am running multiplexed paired-end samples on a HiSeq machine, I can use only 6-nt Craig adapters if I also include a Phix control library? Or is the problem still the same?
I'm confused about this. Basically I am trying to figure out if I can side-step the process of designing my own 4-5nt barcodes since the 6nt ones are already available in the Craig paper. Does anyone know of any references which include sequences for shorter (or longer) barcodes that I could order from IDT, or have tips on how to design them?
I found a thread on multiplexing with Illumina adapters/primers which was very helpful (relevant document attached), as it listed the sequences for Illumina adapters, multiplexing primers, etc. Which ones would be appropriate for paired-end multiplexing? I assume the multiplex ones listed in the document are for single-end reads.
Thanks!
Natalia
Comment