Unexpected lambda spike in distribution

[sgreenb]

I ran a PacBio sequencing run with BstEII digested lambda DNA.

https://www.neb.com/products/n3014-d...%20Information

Since the DNA is digested with a single restriction enzyme, there should be equal moles of each fragment. However, in the run I saw few read counts at the smaller fragment size, then increasing as size got larger, and a slight drop off for the largest fragments.

One of the larger fragments was also a significant outlier, having 1/3 the counts as the next size up and down. This is concerning because it may mean there is some sizing bias in the run. What could be causing this?

[rhall]

Between sample prep and loading of the reads on the machine there are multiple points that introduce bias for fragment size. Standard library prep procedures would select against very short fragments, and loading of fragments for sequencing is generally optimized for long WGS libraries.

[sgreenb]

Between sample prep and loading of the reads on the machine there are multiple points that introduce bias for fragment size. Standard library prep procedures would select against very short fragments, and loading of fragments for sequencing is generally optimized for long WGS libraries.

In figure S3 in the supplemental of this paper the exact opposite trend was seen with PacBio sequencing and the same digested DNA.

https://www.cell.com/molecular-thera...entaryMaterial

Trying to understand why I the opposite trend from this.

[rhall]

The paper uses an RSII instrument and an old chemistry. The distribution would be expected to change between instruments, chemistries and sample prep kits. I would also expect some changes at a lower magnitude to be unique to the specific run. If the intention is to use the BstEII digested lambda to control for the characterization of fragment lengths in the sample then I would suggest that it needs to be included in all runs.