Methylation extraction results

My BAT_calling job finished! It ran for about five days (not continuously) and processed about 4-5 samples with BAT_calling each day. On the last day, four samples were processed with BAT_calling and all samples were processed with BAT_filtering.

BAT_calling output

For each sample, BAT_calling produced VCF files all methylation information in that sample, as well as a log file documenting progress. These logs don’t have any interesting information, but they’re now in this folder. The VCF files are in my home directory on vortex.

BAT_filtering output

The more interesting output from this step in the workflow came from BAT_filtering. I filtered methylation data extracted with BAT_calling for CG methylation only, minimum 10 reads, maximum 100 reads, and a minimum methylation rate of 10%. Filtered CpG information was saved as a VCF and a BEDgraph. This step also produced pdfs for each sample with coverage and methylation information for all CpGs and filtered CpGs. The BEDgraphs and pdfs can be found here. One thing I may want to do when writing these results is to create similar plots for a combined BEDgraph with all my samples, so I can describe the broad methylation landscape. I may also want to get numbers for coverage and methylation rate for each sample instead of relying on the pdfs.

When looking at the coverage plot for filtered CpGs, I noticed that L4 samples had coverage plots that were similar between samples, but different from L2 and L3 samples. I think these may be differences in coverage by lane?

Figures 1-3. Coverage for filtered CpGs for representative L2, L3, and L4 samples.

L4 samples include those from SC hypoxia, and normoxia and outside controls for both populations. The sample that was corrupted was from L4, so there may be something going on with that lane in general. Fewer samples were sequenced on L4 compared to L2 and L4. I looked at mapping statistics for L2 and L3 vs. L4. The only thing that set L4 apart was a lower percentage of mapped multiple pair ends.

Going forward

1. Ask Neel about differences between L4 and L2/L3
2. Review next BAT module
3. Run BAT_analysis