Summary of DNA methylation analysis

A brief summary of our DNA methylation analysis.

RRBS

We used RRBS (reduced representation bisulfite sequencing) to analyse the DNA methylation profiles affected by our experimental feed with different micronutrient compositions. RRBS utilises restriction enzymes to target CpG rich regions in the genome. We used two different enzymes for our RRBS analysis.

Restriction enzymes used for RRBS

  1. MspI (recognition sequence : 5' CCGG)
  2. TaqI (recognition sequence : 5' TCGA)

RRBS samples

We collected 21 liver samples at the final harvest stage for RRBS sequencing.

The number of samples for each group

  • L1 diet: 9 samples
  • L2 diet: 6 samples
  • L3 diet: 6 samples
No Name Diet Sex
1 L1 L2 M
2 L6 L3 M
3 L7 L3 F
4 L11 L1 M
5 L12 L1 F
6 L13 L1 F
7 L14 L1 F
8 L16 L2 F
9 L18 L2 M
10 L20 L2 F
11 L22 L1 M
12 L23 L1 M
13 L24 L1 M
14 L26 L3 M
15 L28 L3 F
16 L30 L3 F
17 L32 L2 M
18 L33 L2 F
19 L38 L3 F
20 L42 L1 F
21 L43 L1 F

Definition of genomic regions

Functions of DNA methylation can be different depending on the types of regions where methylation occurs. We split the genome into seven different regions for our RRBS analysis.

Genomic regions for RRBS read mapping

  1. Exon
  2. Intron
  3. P250 (proximal promoter)
  4. P1K (promoter)
  5. P6K (distal promoter)
  6. Flanks (potential enhancer region)
  7. IGR (intergenic region)
Genomic regions for RRBS read alignment
Definition of genomic regions for RRBS read alignment.

Bioinformatics pipelne for RRBS

We used various bioinformatics algorithms and methods to analyse our RRBS samples. The following tools were those we used in our main RRBS pipeline.

Results

Overall diet effect

Unlike the RNA-seq samples, t-SNE clustering analysis showed no obvious separations of RRBS samples by diet.

t-SNE clustering for RRBS samples
t-SNE (t-distributed stochastic neighbor embedding) clustering of 21 RRBS samples.

Differentially methylated CpG sites

There were no noticeable differences between L2:L1 and L3:L1 as well as hypo- and hyper-methylation in terms of the number of DMCs.

Identified DMCs for L2:L1 and L3:L1

  • L2 vs. L1: 2521 DMCs
  • L3 vs. L1: 2555 DMCs
Violin plots of DMCs
Violin plots of DMCs.

See What are DMCs? for more details about DMCs.

Significantly affected biological pathways

ORA (over representation analysis) on KEGG (Kyoto Encyclopedia of Genes and Genomes) pathways showed that micronutrient supplement significantly affected DNA methylation profiles in cell-adhesion and cell-signalling pathways.

Enriched KEGG pathways by ORA (over representation analysis).
Dataset Pathway KEGG ID ORA (gene ratio)
L2:L1 Cell adhesion molecules (CAMs) sasa04514 RS+GB (38/616), P+GB (32/517), Gene body (31/420), Intron (26/339)
Glycosaminoglycan degradation sasa00531 P1K (2/21)
Mitophagy - animal sasa04137 Promoter (7/99), P1K (3/21)
Apelin signalling pathway sasa04371 P+GB (28/517)
Starch and sucrose metabolism sasa00500 P6K (4/74)
L3:L1 Cell adhesion molecules (CAMs) sasa04514 RS+GB (37/573), P+GB (30/465), Gene body (27/388), Intron (27/320)
ECM-receptor interaction sasa04512 Intron (15/320)

Page links

Overview

DMC: differentially methylated CpG site

DMG: differentially methylated gene

Leave a comment