Tweet
We have done 4 runs of transcriptomes with size selection, and the read length is not like it should be.
1. Our first few runs were without size selection, and the sequencing looks fine. We had the same sizes on the sequencer as on the Bioanalyzer.
2. Because we are losing a lot of possible data with the short reads, we did a size selection of 200 bp on the above sample.
The size selected material is from the tip of the tail of the Gaussian cuve, and not from the middle.
Sequencing of this gives bad results : very short readlengths, most below 100 bp, size selection was done at 200 bp. We repeated this for an other sample with the same results.
3. We did this for a new sample, size selection around 250 bp, this time, the tail of the curve went up to 500 bp, so the sample was still from the tail, but not the very end.
sequencing was disappointing again (but better than the previous one, most reads below 150 bp) :
4. To prove that sequencing of a 250 bp fragment is possible with the 200 bp kit, we did it also starting from gDNA, and that worked fine, most reads around 250 bp.
So my question is : why does the sequencing works fine with DNA, and not so good with cDNA from transcriptomics ? Is the size-selected material from the tail problematic cDNA (structure, base sequence, repeats, ...).
Anyone seen the same thing with size selected cDNA from the middle of the Gaussian cuve ?
Best regards,
Andy
1. Our first few runs were without size selection, and the sequencing looks fine. We had the same sizes on the sequencer as on the Bioanalyzer.
2. Because we are losing a lot of possible data with the short reads, we did a size selection of 200 bp on the above sample.
The size selected material is from the tip of the tail of the Gaussian cuve, and not from the middle.
Sequencing of this gives bad results : very short readlengths, most below 100 bp, size selection was done at 200 bp. We repeated this for an other sample with the same results.
3. We did this for a new sample, size selection around 250 bp, this time, the tail of the curve went up to 500 bp, so the sample was still from the tail, but not the very end.
sequencing was disappointing again (but better than the previous one, most reads below 150 bp) :
4. To prove that sequencing of a 250 bp fragment is possible with the 200 bp kit, we did it also starting from gDNA, and that worked fine, most reads around 250 bp.
So my question is : why does the sequencing works fine with DNA, and not so good with cDNA from transcriptomics ? Is the size-selected material from the tail problematic cDNA (structure, base sequence, repeats, ...).
Anyone seen the same thing with size selected cDNA from the middle of the Gaussian cuve ?
Best regards,
Andy
Comment