Shorter fragments than expected

[peripoe]

My problem is having shorter fragments after sequencing than expected. The size of my library should be around 300 bp and I checked the size with Tape station which was ~300 bp. However, my sequencing results show that my actual transcripts are too short that that even can`t be mapped to anywhere. Then, I checked my libraries with Sanger sequencing and they were also around 150 bp.
What can it be the reason that I see at Tape station results that 300-350 bp but our fragments are less than 150??
The only mistake I am sure is my libraries (CEL-Seq libraries, which you can pool many samples in one tube) is not compatible with Hi-Seq v4 chemistry (don`t know the reason but the method published to be so, now the authors says it works with v4 only %20 Phix is used) but works fine with v3 chemistry or Rapid mode. Previously the same protocol worked fine with Rapid mode and now we tried v4 with %1 Phix and the results are very strange with short fragments.
Any idea??

[nucacidhunter]

Two possible reasons:
1- Tape Station indicates library fragments length which includes insert plus adapter. Average adapter length for standard libraries are around 120b but Cell-seq adapters are longer due to inline cell specific barcodes so the transcript insert length are shorter
2- Shorter fragments in libraries are sequenced preferentially so average sequence read length will be shorter than average library insert

You have not mentioned sequencing configuration. Cell-seq design is poor as in standard sequencing R1 has to be read at least 25 cycles. This will include reading through polyT and without high PhiX spike-in the %PF will be low.

Posting library Tape profile and FastQC results will give better ideas for troubleshooting.

[peripoe]

Hi,

Please see the files that I upload. I also need to correct that Sanger result is 180 bp including adapters which leaves 50-60 bp and sometimes shorter fragments. With Tape station, I see a peak ~300 bp which makes us think that our product is ~150 bp.

[nucacidhunter]

Tape profile shows some adapter/primer dimers that they also will be sequenced preferentially along with shorter insert fragments.

Plots confirm sequencing short inserts as after cycle 80 polyA is clearly distinguishable and overall high A residue also indicates sequencing short fragments containing only polyA. If you need longer reads for mapping then fragmentation step has to be modified to get larger RNA if you are following original publication protocol. This can be done by one way bead size-selection on the current library as well but I am not sure what will be the effect on results.