Skip to content

Latest commit

 

History

History

medipseq_replication

README.md

medipseq_replication

Background

This code runs analysis for the paper "Genetic variation influencing DNA methylation provides new insights into the molecular pathways regulating genomic function" by Hawe et al. Specifically, it runs the analysis for replicating the HSM-LD block associations from the MeDIP-seq results of Bell et al. (PMID: 29295990).

Code

The R scripts are found in 'code/' and are the two files:

  • step_1_map_HSM_LD_assoc_to_SNP_CpG.r

We take the HSM-LD block associations from PMID: 29295990 and map the associations to pairs of SNPs and CpGs analyzed in our data.

  • step_2_calculate_enrichment_above_bg.R

Based on summary statistics, we see how many of the HSM-LD block associations we can replicate with pairs of SNPs and CpGs in our data. We check also the summary statistics for the 100 sets of matched SNP-CpG pairs (background) to see if the HSM-LD-mapped SNP-CpG pairs show a greater number of significant hits than the background.

Data

The data required for the code are found at 10.5281/zenodo.5196216 and are the following:

  • mergeFinal_hsm_LDblock_GWASSNPs.txt

Results from PMID: 29295990.

  • illumina_450K_annot_pruned.txt

Pruned Illumina annotation file.

  • kora_cpgs.RData

CpG sites used in our analysis in KORA (ie, passing QC).

  • kora_snps.RData

SNPs used in our analysis in KORA (ie, passing QC).

  • medip_replication_results.txt

The linear regression (association) results of the HSM-LD-mapped SNP-CpG pairs.

  • medip_matched_background_results.txt

The linear regression (association) results of the 100 sets of matched SNP-CpG pairs (background).

Notes

It is possible to completely reconstruct the results from the data found here. However, the data presented here are not raw data. Specifically, after step_1, the following cannot be performed due to data privacy reasons:

  • we mapped the CpG-SNP pairs generated in step_1 to 100 sets of matched background CpG-SNP pairs
  • for the the CpG-SNPs pairs and then their matched sets, we used individual-level KORA F4 data to check the associations, using linear regression

The results of these analyses (medip_replication_results.txt and medip_matched_background_results.txt) are then used as inputs for step_2.