Creating Beamer slides with optionally included notes slides using R Markdown

Using RStudio, it is very easy to create Beamer slides using R Markdown.  RStudio also supports LaTeX, so this enables even more flexibility.

Using the template shared here:

https://github.com/DanielEWeeks/Beamer-Rmd-Notes-Slides

one can also insert 'note slides' that can easily be all included or all excluded from the generated set of slides, just by toggling one word in the LaTeX header.

This allows one to use a more concise set of slides, without the notes, in class while providing the students a more comprehensive set that contains all the notes.

It could also be used to hide 'answer' slides from the students' slide set when posing questions in class.  When posing questions in class, I prefer that the students don't have the 'answer' slides right in front of them.  Then, after class, I make the slide set that contains the 'answer' slides available.

I previously shared instructions on a different approach for doing this when generating Beamer slides using LaTeX or LyX - see:

http://deweeks.blogspot.com/2014/08/hiding-answer-slides-in-student.html

Optical character recognition

Here is a script that uses the 'tesseract' optical character recognition software to extract recognizable text from a PDF file:

#!/bin/bash

# Purpose: 

#   To carry out OCR on a PDF file

if [[ $# -ne 1 ]]; then

  echo "This script expects one argument."

  echo "  This argument is the name of the pdf file" 

  echo "  including the .pdf extension"

  echo "Usage: $0 file.pdf "

else

filename=$(basename "$1")

filename="${filename%.*}"

echo "Converting $1 to a tiff file named $filename.tiff"

echo "... "

convert -density 300 $1 -depth 8 -strip -background white -alpha off $filename.tiff

echo "Carrying out OCR on $filename.tiff to create $filename.txt"

tesseract $filename.tiff $filename

echo "The recognizable text in $1 has been output to the $filename.txt file."

fi

For information on how to install the needed software, see this web page.

Please assign informative names to downloaded PDFs!

Bravo for the journals that assign human-readable informative names to the PDF versions of articles downloaded from their web sites.  For example, these naming schemes are very nice:

Hum. Mol. Genet.-2015-Simpkin-3752-63.pdf
PNAS-2005-Storey-12837-42.pdf
Int. J. Epidemiol.-2015-Sharp-1288-304.pdf


Boo for the journals that assign non-informative names to their PDF files.  For example, these naming schemes are not informative to me (even though the DOI is part of the name of the first two):

art%3A10.1186%2Fgb-2013-14-5-r42.pdf
art%3A10.1007%2Fs11357-016-9927-9.pdf
ijerph-12-14461.pdf

Simple parallelization using 'sem' of the GNU parallel package

If you have a multiprocessor computer, you can easily use the ‘sem’ part of the GNU Parallel system to run processes in parallel.   

It was really simple to use and worked as intended.  And this was much easier to install and get working than the other approach I was contemplating, which was installing grid engine software.

See:

https://www.gnu.org/software/parallel/sem.html

Below is my parallelized script, which ran a whole set of “HaploPS” commands in parallel using 12 of my processors.  I have bolded the two 'sem' commands that made this script execute in parallel.

$ more run_haploPS_parallel.sh 

#!/bin/bash

for (( chr=1;chr<=22;chr=chr+1)) {

     cd $chr

     for (( i=95 ; i>=5 ; i=i-5)) {

        freq=$(echo "scale=2;$i/100"|bc)

        date

        echo "Starting HaploPS run on chromosome " $chr " with freq = " $freq

        sem -j12 HaploPS -geno selscan.hap -legend selscan.map -freq 0$freq -out ../haploPS/haploPS_0${freq}_chr${chr}.txt 

        echo "Run completed."

        date

}

      cd ..

}

sem --wait

##############It will automatically run haploPS at 5 to 95 percent frequencies.

#################################################################

The difference between mathematics and statistics

"Mathematics is about whether the conclusions follow from the assumptions. By contrast, statistics is about whether the assumptions have anything to do with the real world." 

Jay Kadane, Leonard J. Savage Professor of Statistics, Emeritus, Carnegie-Mellon University

rsync

As part of transitioning from using an old desktop to a new desktop (while both are still active), I found the 'rsync' command to be very useful.  Here is the options I ended up using to accomplish my goal of copying over (old) files from the old machine to the new without wiping out any new files on the new machine:

rsync -avzuhP --log-file=/destination/dir/rsync_log.txt -e ssh remoteuser@remotehost:/source/dir /destination/dir/

Here's what the options do:
-a archive
-v verbose
-z compress
-u update  <= Do Not Overwrite the Modified Files at the Destination
-h human-readable
-P progress bar/partial transfers

Note that these options are turned on in '-a' archive mode:
        -r, --recursive             recurse into directories
        -l, --links                 copy symlinks as symlinks
        -p, --perms                 preserve permissions
        -t, --times                 preserve times
        -g, --group                 preserve group
        -o, --owner                 preserve owner (super-user only)
        -D                          same as --devices --specials

Hiding 'answer' slides in student handouts using Beamer

After adding 'Question' slides, followed by 'Answer' slides, to my Beamer presentation, I wondered if there was an easy way to remove the 'Answer' slides from the handout version of the slides that I will give to the students.

It turns out that there is, as described in this link:  just add <handout:0> right after the \begin{frame}

When 'handout' is added to the document class, all those slides marked with <handout:0> are automatically excluded.

This is wonderful!

The wonderful 'endfloat' LaTeX package

With LaTeX (or LyX), one can work on a manuscript while placing the tables and figures in their natural places, and then, with the addition of a single line to the preamble:

\usepackage[nolists,tablesfirst]{endfloat}

the tables and figures are magically moved to the end, as is required at submission by many journals.  

The 'endfloat' package even inserts markers like "[Table 1 about here]" in the main text.

Now I just need to persuade my colleagues to start writing their manuscripts in LaTeX...

Here is an amazing set of photographs of C.C. Li, who was the founder of the Division of Human Genetics within the Department of Biostatistics at the University of Pittsburgh.  

Markov Chain Humor

Possible mascots for a program that uses Markov Chains in its computations:

Markov Man
   "Look, it's a bird... no, it's a plane,... no, I forgot what it was."


Hidden Markov Man
   the truly Forgetful Man

Computers are useful?!

"A computer makes it possible to do, in half an hour, tasks which were completely unnecessary to do before" ~ Dave Barry.

Zero p-values in R

For the top hits from an R program, I was getting all zero p-values, which are not that useful for ranking or plotting.

Turns out that the R program was calculating the p-value in this manner:

1 - pchisq(71.12830,1)

But, according to the discussion here, in R

".Machine$double.eps is the smallest number such that 1+x can be distinguished from 1"

On my machine, we have that:

> .Machine$double.eps
[1] 2.220446e-16

So that is why this truncates to zero:

> 1 - pchisq(71.12830,1)
[1] 0

However, we can get a non-zero p-value if instead we compute it this way:

> pchisq(71.12830,1,lower.tail=FALSE)
[1] 3.347341e-17

knitr

I recently became aware of the kintr package http://yihui.name/knitr/ which is much more capable and easier to use than Sweave when creating dynamic 'reproducible research' reports that interweave LaTeX and R.

Quertle http://www.quertle.info/

I learned today about an amazing new program for searching the PubMed literature database plus full text documents - this is called 'Quertle'  http://www.quertle.info/ 

I tried it out and quickly found an important and relevant paper that we hadn't found in an earlier more extensive conventional literature search.

Your genetic profile

Craig Venter said, in response to a question about the 'risks' of knowing one's genetic makeup:

"Understanding the genetic code is understanding probabilities.  There is very little in your code that is yes/no."

Hazards of Having Good Students in your Introductory Statistics Class

This paper "Hazards of Having Good Students in your Introductory Statistics Class" by Trivellore Raghunathan about the meaning of confidence intervals and p-values is well-written and entertaining.

The meaning of p-values

Regarding p-values, Paul Meehl said something along the lines of "a p-value answers a question we shouldn't ask about a topic we aren't interested in".

THINK

"I've often said that once scientists get a p-value less than 0.05 or a confidence interval that implies statistical significance, they turn their brains off."  Steve Simon, quoted from here.

Significance

This cartoon from XKCD is wonderful:

"Any claim coming from an observational study is most likely to be wrong"

This article from the journal "Significance"

Deming, data and observational studies
Author: S. Stanley Young, Alan Karr
Published: Aug 25, 2011 - From issue: Volume 8 Issue 3 (September 2011)
Doi: 10.1111/j.1740-9713.2011.00506.x

suggests a process control-based solution for improving the success rate of observational studies.

To better appreciate the problem of multiple testing, they suggest visiting Jery Dallal's simulation web page, which illustrates what happens when one carries out "100 Independent 0.05 Level Tests For An Effect Where None Is Present".