dss-snakemake

A Bioinformatics pipeline for differential methylation analysis using DSS. Pipeline is WIP!

Getting started

1. Clone the repo
2. Install Snakemake version >= 7.6.0
3. Fill in the config and manifests.
4. Start the analysis!
   • Begin with a dry-run

   ./runlocal 10 -np
   

   • If dry-run looks good, proceed with:

   ./runlocal 10
   

Pre-reqs

config/
└── config.yaml

Config explanation

mode: group # group or regression model- only group is supported at the moment.
dss_cntr: "/path/to/singularity.sif"
gtarget: autosome # autosome, sex, or wgs
tech: ont # hifi or ont

manifest: config/manifest.tab # tab delimited columns: sample_name, bed, group

dss_exclude_regions: "/path/to/lc.bed.gz" # the pipeline only parses the three columns, bed3
dss_min_mod: 0 # optional- if you want to exclude any sites with zero methylation

# optional annotations but I do parse the keyword genes special as in I get the distance from the gene as well as intersection
annotations:
    cpg_island: ""
    genes: ""

# optional but please make sure all keys are strings
# each key (e.g. one) corresponds to the name in the manifest under group and purpose explained in next section 
group_target:
    "one": "sample_child"
    "another_one": "sample_child_hap2;sample_child_hap1" # another example of if you want additional target values.

The purpose group target is to target the sample within a group and exclude any pairs that does not contain target sample. For example:

$ cat config/manifest.tab
sample_name	bed	group
sample_father	sample_father_cpg-pileup.bed.gz	one
sample_mother	sample_mother_cpg-pileup.bed.gz	one
sample_child	sample_child_cpg-pileup.bed.gz	one

gives us, the following pairs:

results/ont/dss/group_comparison/one/autosome/
├── sample_father_vs_sample_child_DMR-prcnt_summary.tsv.gz
└── sample_mother_vs_sample_child_DMR-prcnt_summary.tsv.gz

and NOT

├── sample_father_vs_sample_mother_DMR-prcnt_summary.tsv.gz

FAQ

1. What do the results mean?
   1. sample_father_vs_sample_child_DMR-prcnt_summary.tsv.gz tells us the differential methylated regions when we compare father and child. In other words, relative to the father- these are the regions that are differential. Please refer to the DSS manual for output explanation.
2. Must I have the keyword genes?
   1. If you're providing a bed file of genes, please use the key "genes" to get both intersection and distance to that intersection.