Switch from colMeans() to colMeans2() in preprocessNoob incurs numerical differences

Switching colMeans() to colMeans2() is tempting; it will be more efficient because we don't allocate the intermediate matrix.

https://github.com/hansenlab/minfi/blob/a790afc1a5c878d235f9e91fd604288c5411c443/R/preprocessNoob.R#L123-L124

https://github.com/hansenlab/minfi/blob/c6d40d432504703d1d5a0397bd5f9e2e690a57ae/R/preprocessNoob.R#L89-L94

However, the results from using colMeans() vs. colMeans2() are not numerically identical!

library(matrixStats)
set.seed(666)
x <- matrix(runif(1000, 100, 1000000), ncol = 10)
colMeans(x) - colMeans2(x)
#>  [1]  1.746230e-10  1.746230e-10  0.000000e+00  5.820766e-11  1.164153e-10
#>  [6]  5.820766e-11 -5.820766e-11 -2.910383e-10 -3.492460e-10 -1.746230e-10

Created on 2018-04-18 by the reprex package (v0.2.0).

Session info

``` r devtools::session_info() #> Session info ------------------------------------------------------------- #> setting value #> version R version 3.5.0 RC (2018-04-15 r74605) #> system x86_64, darwin15.6.0 #> ui X11 #> language (EN) #> collate en_AU.UTF-8 #> tz America/New_York #> date 2018-04-18 #> Packages ----------------------------------------------------------------- #> package * version date source #> backports 1.1.2 2017-12-13 CRAN (R 3.5.0) #> base * 3.5.0 2018-04-16 local #> compiler 3.5.0 2018-04-16 local #> datasets * 3.5.0 2018-04-16 local #> devtools 1.13.5 2018-02-18 CRAN (R 3.5.0) #> digest 0.6.15 2018-01-28 CRAN (R 3.5.0) #> evaluate 0.10.1 2017-06-24 CRAN (R 3.5.0) #> graphics * 3.5.0 2018-04-16 local #> grDevices * 3.5.0 2018-04-16 local #> htmltools 0.3.6 2017-04-28 CRAN (R 3.5.0) #> knitr 1.20 2018-02-20 CRAN (R 3.5.0) #> magrittr 1.5 2014-11-22 CRAN (R 3.5.0) #> matrixStats * 0.53.1 2018-02-11 CRAN (R 3.5.0) #> memoise 1.1.0 2017-04-21 CRAN (R 3.5.0) #> methods * 3.5.0 2018-04-16 local #> Rcpp 0.12.16 2018-03-13 CRAN (R 3.5.0) #> rmarkdown 1.9 2018-03-01 CRAN (R 3.5.0) #> rprojroot 1.3-2 2018-01-03 CRAN (R 3.5.0) #> stats * 3.5.0 2018-04-16 local #> stringi 1.1.7 2018-03-12 CRAN (R 3.5.0) #> stringr 1.3.0 2018-02-19 CRAN (R 3.5.0) #> tools 3.5.0 2018-04-16 local #> utils * 3.5.0 2018-04-16 local #> withr 2.1.2 2018-03-15 CRAN (R 3.5.0) #> yaml 2.1.18 2018-03-08 CRAN (R 3.5.0) ```

For further details, see https://github.com/HenrikBengtsson/matrixStats/issues/96

Consequently, switching to colMeans2() will break the digest-based unit tests.

hansenlab / minfi

Switch from colMeans() to colMeans2() in preprocessNoob incurs numerical differences #153