Journal club 20201026

count: false
class: left, bottom

# Journal club 
## October 26, 2020

<img src="IEU-logo-colour.png" width="50%">
---
layout: true

.logo[.mrcieu[
MRC Integrative Epidemiology Unit    
]]

---

## EWAS of disease

---

## EWAS of exposure

.striped[
|     PMID|Title                                                                                                                                                     |Journal           |Variable                         |Tissue                                                         |n                  |Assocs                                     |Note                                                                                  |
|--------:|:---------------------------------------------------------------------------------------------------------------------------------------------------------|:-----------------|:------------------------------------------|:--------------------------------------------------------------|:----------------------------|:------------------------------------------------|:--------------------------------------------------------------------------------------------|
| 32916427|An epigenome-wide association study ... of exposure to endocrine disruptors.                                      |Environ Int       |13 endocrine disrupting chemicals in urine |Blood                                                          |622                          |4 CpG sites                                      |                                                                                             |
| 32928285|DNA methylation mediates the effect of cocaine use on HIV severity.                                                                                       |Clin Epigenetics  |Cocaine use                                |Blood                                                          |1435                         |15 CpG sites                                     |3/15 mediators by 2-step MR                                   |
| 33079621|Birth weight associations with DNA methylation differences in an adult population.                                                                        |Epigenetics       |Birthweight                                |Blood (adult)                                                  |1757                         |1 CpG site                                       |                                                                                             |
| 32894192|Associations of maternal early-pregnancy blood glucose and insulin concentrations with DNA methylation in newborns.                                       |Clin Epigenetics  |Maternal blood glucose and insulin         |Cord blood                                                     |935                          |Null                                             |                                                                                             |
| 33023569|... maternal tobacco consumption alter placental DNA methylation ...          |BMC Med           |Prenatal smoking                           |Placenta                                                       |568                          |242 CpG sites                                    |                                                                                             |
]

---

## EWAS of germ cells

However, also just published ...

.box[
Åsenius, et al. *The DNA methylome of human sperm is distinct from blood with little evidence for tissue-consistent obesity associations.* PLoS Genet. 2020 Oct 13;16(10):e1009035.

* **Only** "one cross-tissue obesity-specific hypermethylated site (cg19357369; chr4:2429884; P = 8.95 × 10-8; 2% DNA methylation difference) was identified, warranting replication and further investigation."

* **Therefore** "spermatozoal DNA methylation is an unlikely mediator of intergenerational effects of metabolic traits."
]

---

## EWAS of phenotype

---

## EWAS for prediction

---

<img src="methhaplo.png" width="70%">
--
.rightmid[
.box[
* **Method** "uses allele-specific DNA methylation and SNPs from WGBS"

* **Result** "methylation haplotype regions were ten times longer than haplotypes with SNPs" 
]]

---

<img src="capture.png" width="70%" >
--
.rightmid[
.box[
TruSeq EPIC 'better' than EPIC-array:

* wider dynamic range and better precision for hypo- or hyper-methylated DNA.

* identified more differentially methylated CpG sites
]
]
---

.iii[
<img src="pnas.jpg" width="90%">

*Adding timing/development to GxE studies.*

https://www.pnas.org/content/117/38
]
.gap[ ]
.iii[
<img src="pnas-intro.png" width="100%" >
<img src="pnas-perspectives.png" width="100%" >
<img src="pnas-research1.png" width="100%">
]
.iii[
<img src="pnas-research2.png" width="100%">
<img src="pnas-research3.png" width="100%">
]

---

## Mitochondrial DNA copy number outcomes may be mediated by DNA methylation

> Castellani CA, Longchamps RJ, Sumpter JA, et al. 
> **Mitochondrial DNA copy number can influence mortality and cardiovascular disease via methylation of nuclear DNA CpGs.** 
> Genome Med. 2020;12(1):84.

### Background

* mtDNA-CN is tissue-specific

* lower CN is bad!

*(altered cell morphology,
    lower respiratory activity,
    CVD,
    chronic kidney disease,
    diabetes,
    liver disease,
    fraitly,
    all-cause mortality)*
--

* lower CN associated with cancer progression via DNA hypermethylation 
    and changes in gene expression

---
layout: true

## Mitochondrial DNA copy number, continued

.logo[.mrcieu[
MRC Integrative Epidemiology Unit    
]]
---

### Result 1: mtDNA-CN associated with DNA methylation

* DNA methylation (n=2507, ARIC study)

* Associated with 34 CpG sites

* Enrichments: cell signaling and the “neuroactive ligand-receptor interaction” pathway

* 7 of 34 associations replicated (n=2528)

---

### Result 2: *TFAM* knockout generates similar DNA methylation changes

* *TFAM* is a well-known regulator of mtDNA replication

* Knockout in HEK293T cells by CRISPR-Cas9

* 5-fold reduction in expression and 18-fold reduction in mtDNA CN

* Same pathway (*neuroactive ligand-receptor interaction pathway*) enriched in both 
    top differentially methylated CpG sites and differentially expressed genes

* 3 of 7 replicated sites differentially methylated (p < 0.05)

---

### Result 3: DNAm changes do not influence mtDNA CN

* Standard Mendelian Randomization within discovery dataset (ARIC; n=2507)

* Null (all p > 0.1 for Bonferroni adjusted threshold p < 0.005)

* Power for individual CpG sites was low. *Could have used 2-sample MR?*

---

### Result 4: mtDNA CN-associated CpG sites associated with outcomes

* Considered 6 of 7 replicated CpG site associations

'`*`' association after adjustment for multiple tests