Comparing Bisulfite Sequencing Approaches

We took DNA from a single Eastern oyster and prepped using MBD, MSP digestion, and plain old DNA. Full details can be found here.

Five samples were sent to Qiagen for library prep and sequencing.

gDNA C.virginica – Genomic DNA, 20uL, 58.4ng/uL
MBD 1 virginica – Fragmented (~400bp average size), MBD-enriched, 25uL, 18.3ng/uL
MBD 2 virginica – Fragmented (~400bp average size), MBD-enriched, 25uL, 19.6ng/uL
MspI 1 virginica – gDNA digested with MspI, 25uL, 53.4ng/uL
MspI 2 virginica – gDNA digested with MspI, 25uL, 31.0ng/uL

From this, 20 preps were made. libraries

With QC on these files available.

align

Some libraries had a low output and were dropped. To do a comparison Bismark was performed looking at 100k reads.

The gDNA and MSP libraries had methylation calls (no filters) at about 7% whereas the MBD libraries hit at about 70%.

meth

None of this is too surprising. What is of most interest is for any given loci (CpG), what is the coverage and where are the loci. As for coverage given 100k reads…

Library 3x 10x
01-gDNA 742 17
07-msp 5153 227
08-msp 3943 169
09-MBD 1170 10
10-MBD 1037 8
11-MBD 496 5
12-MBD 582 7
13-MBD 873 9
14-MBD 1146 22
15-MBD 807 12
16-MBD 741 8
17-gDNA 27 0
18-gDNA 68 0
19-gDNA 98 0
20-MBD 813 8

IGV

igv

You can play with yourself with loading this session:

https://raw.githubusercontent.com/sr320/nb-2018/master/C_virginica/igv/Qiagen-0915-igv_session.xml

igvmov

There a few more questions that could be tackled. Post in the comments if something comes to mind.

Written on September 15, 2018