Hi!
I received some data produced by single-cell sequencing (smart-seq2). After aligning onto the corresponding genome (+ERCC) I got a matrix of count which look quite sparsed:
ID start end strand length count1 count2 count3
ERCC-00004 1 523 + 523 1902 145 2328
ERCC-00009 1 984 + 984 318 29 428
ERCC-00012 1 994 + 994 0 0 0
ERCC-00013 1 808 + 808 0 0 0
ERCC-00014 1 1957 + 1957 0 0 0
With typically higher concentration ERCCs being detected but not lower ones.
I know for sure that those samples were not sequenced deep enough but I wanted to know if for ERCCs in single-cell sequencing I should expect to see every ERCC detected or is it common to see some drop-off ?
Thanks in advance!
•
written
by
VHahaut
The ERCC RNA spike-in mixes manufactured by Thermo Fisher cover six orders of magnitude of concentrations, therefore it is expected by design that a large number of them will not be detected. You can have a look at Svensson et al., 2007 for an example on how to analyse the spike data.
Side comment: while people have been for a long time aligning their reads to the sequence of the plasmid inserts used to produce the spikes, the NIST now distributes the sequence of the transcribed products, which are more accurate at the 5′ end. It may not make a big difference for a RNA-seq analysis, but I recommend to use these reference sequences (disclaimer: I contributed to prepare the file, see Patch ERCC spike sequences to get their real 5-prime ends. for details).
Isn't ERCC-00012 one of the ERCC controls with amongst the highest concentrations, though? I would have expected that to show some counts.
Then it makes sense what Charles wrote. Great, thank for confirming.
Log in to answer this question.
Hi! I also received Smart-seq2 data. I never analyzed data coming from... so I would ask how to perform the first steps. I got XX fastq files.. so I first trimmed after a QC report then I would align them separately with star or tophat and then quantify the abundance of each transcript with RSEM or HTseq count.. then use R (scater) to perfom further analysis.. is this workflow correct or should I use different tools ??
thanks in advance