Wednesday, June 28, 2017

Intelligence & genetics

A news-item on GenomeWeb:

Rather than having genetic variants that make them smart, brainy people may lack mutations that make them less clever, New Scientist reports.

In a paper posted to BioRxiv, researchers from the University of Edinburgh report that they genotyped some 20,000 people from the Generation Scotland family cohort to tease out the effects of gene variants on intelligence, extraversion, and neuroticism. As the cohort includes family members, the researchers could delve into variants not typically found in genome-wide association studies of unrelated people.

.....
In particular, New Scientist says that CNVs, structural variants, and rare variants seem to affect intelligence. As rare variants are more likely to be harmful, New Scientist says it appears that a person's intelligence might be in part due to their mutational load.
Note 1: CNV = copy number variations
Note 2: It would seem high intelligence is the norm, and variation away from the norm reduces intelligence.  Rather amusing, and this would be a blow to the IQ-metricians, I think.

Thursday, June 22, 2017

Grey Catbird

New to my yard, as far as I can remember. I believe it is a grey catbird.





Monday, June 05, 2017

Acceleration in the rate of sea level rise

Some people are looking at sea level data and trying to fit a quadratic time curve to find the alleged acceleration in the rate of rise of sea level.   This doesn't work, and so they are very skeptical that the rate of rise of sea level has accelerated over the last century.

IMO, really they should be looking for piecewise linear fits, and a change in slope of the line segments. The point is that acceleration is simply a change in a rate; and nobody has claimed a constant acceleration in the rate of sea level rise. The rate of sea level rise has changed.

E.g., using data and the graphing utility at sealevel.info, here are three graphs.

Here is sea level data at Delfiziji, Netherlands from January1865 to December 2015.  The fitted line has a slope of 1.72 +/- 0.14 mm/year.

1/1865 - 12/2015 - 1.72 +/- 0.14 mm/yr
The web page also gives this:
Regressions
Linear:
  y = B + M·x
  y = 6797.353 + 1.716·x mm
Quadratic:
  y = B' + M·x + A·x²
  y = 6784.346 + 1.716·x + 0.00685·x² mm
where:
  Date range = 1865/1 to 2015/12
  x = (date - 1940.46(i.e., 1940/6)
  slope = M = 1.716 ±0.141 mm/yr
  acceleration = 2·A = 2×0.00685 = 0.01369 ±0.00722 mm/yr²
The posited constant acceleration is there, but extremely tiny and buried in the noise.

We now break the time series into two periods of about 75 years each, from January 1865 to December 1940; and January 1941 to December 2015.
The first period has the sea level rising at 1.31 +/- 0.37 mm/year.
The second period has the sea level rising at 2.25 +/- 0.43 mm/year.

Note that the confidence intervals don't overlap (i.e., 1.31 + 0.37 = 1.68; 2.25 - 0.43 = 1.82).
The rate of sea level rise has increased, ergo, accelerated.
1/1865-12/1940 - 1.31 +/- 0.37 mm/yr

1/1941 - 12/2015 - 2.25 +/- 0.43 mm/yr