Read on Twitter
1/8. Scars of childhood socioeconomic stress: A systematic review - https://tinyurl.com/y43mtgpd
I love that 1st PhD paper feeling we get for our students. This reflects the hard work & persistence of @natashamwood15!
(w/ @ThomasTrebilco @NurtureOfNature)
...So what did we find?...
I love that 1st PhD paper feeling we get for our students. This reflects the hard work & persistence of @natashamwood15!
(w/ @ThomasTrebilco @NurtureOfNature)
...So what did we find?...
2/8. While there is some evidence for some variation in DNA methylation associated with socioeconomic status, class, or position, no firm conclusions can be drawn...
....Why is that?....
....Why is that?....
3/8. We have a problem with our initial definition and operationalisation of socioeconomic status, class, or position... as indicated by the fact we don't even have a single name for it....!
4/8. Some use single components (financial/educational/occupational), others use a random mix. Our group won't be innocent either, restricted by data in cohorts already collected...In the future we need consistently collected cohorts w/ reliable & valid socioeconomic measures...
5,8. Step 1: we need a clearer understanding of what SES/SEP/SEC is, and how to measure it.
But when we look at DNA methylation we stray into another array of challenges... candidate genes can be problematic, and often don't control for critical factors (cell count & pop strat.)
But when we look at DNA methylation we stray into another array of challenges... candidate genes can be problematic, and often don't control for critical factors (cell count & pop strat.)
6,8. Epigenome studies also do not always control for, or report, necessary factors... & as always, power is an issue - confounded by a noisy predictor variable!
We need more consistency in our predictor &in the reporting of our outcomes -when we can't be, we must be OPEN on it.
We need more consistency in our predictor &in the reporting of our outcomes -when we can't be, we must be OPEN on it.
7,8. But the future is bright. Be aware when constructing SEP measures:
- do different subtypes have different effects?
- individual measure of SEP does not map well to population/group SEP (eg., postcode).
- are there critical periods?
- does changing SEP have impact?
- do different subtypes have different effects?
- individual measure of SEP does not map well to population/group SEP (eg., postcode).
- are there critical periods?
- does changing SEP have impact?
8,8. With DNA methylation:
-have you controlled what you can control for?
-if not have you declared that and why?
-are your conclusions couched in that?
-noisy predictor will make it harder to detect an effect!
-pre-registration is your friend!
Thank you for reaching the end!
-have you controlled what you can control for?
-if not have you declared that and why?
-are your conclusions couched in that?
-noisy predictor will make it harder to detect an effect!
-pre-registration is your friend!
Thank you for reaching the end!
Oh - one more, because you can't edit tweets - I am still learning!
I should have tagged my lab in this @BehGEMs whoops!
I should have tagged my lab in this @BehGEMs whoops!
