Initial commit

The file "draft_detailled_protocol_and_booklet_description.md" contains a
summary of the tutorial for the booklet as well as the course protocols.

It also contains background on the data and pdfs of a paper where a data strategy
similar to the one employed by Thomas was used.

.gitignore

+#.gitignore
+131022014_Detailed_Protocols.docx
+Booklet_2016.docx
+#"Phenotypic profiling of the human genome\nby time-lapse microscopy reveals cell\ndivision genes-supplement.pdf"
+#"Phenotypic profiling of the human genome\nby time-lapse microscopy reveals cell\ndivision genes.pdf"
+#draft_detailled_protocol_and_booklet_description.md
+microscopy_Thomas.md
+commit_msg.txt

draft_detailled_protocol_and_booklet_description.md

+
+# Short description of the practical
+
+P12	Data handling and visual exploration of the processed and scored data from P3 / P9
+(ATC, Flex Lab B)
+
+
+Staff: Andrzej Oles, Mike Smith, Bernd Klaus
+
+Work:
+1. Importing the data into R, the concept of "tidy data", introduction 
+to data handling strategies 
+2. Visual exploration of data using quality-control related plots, e.g. heatmaps,
+and PCA plots
+
+
+
+
+# Short description of the practical
+
+
+High-throughput microscopy screens with technologies such as RNAi, CRISPR-Cas
+and libraries of drug compounds typically generate large quantities of data that
+are potentially rich in biological information. Typically thousands of gene or
+drug targets are screened and tens or even hundreds of image features are
+extracted. Exploring these large datasets is challenging. R packages from
+the "tidyverse" will lead the way here.
+
+
+In the tutorial we will fist introduce the concept of "tidy data" that provides
+a practically useful way of organizing big experimental datasets
+and show how turning the initial data in to a "tidy" representation.
+
+We will then demonstrate how to perform large-scale visualization of the screen 
+results and show how to use this to explore patterns in the data. 
+
+Methods such as Principal component analysis (PCA)  and clustering will be
+employed and the participants will be introduced to the advanced graphical 
+capabilities of R.
+
+We will work on the cell-classification results from labs P3 / P9, so this 
+this lab will lead the way to hit calling
+strategies and comprehensive downstream analyses.
+
+We will perform all analyses using open source R/Bioconductor software. 
+
+
+Software:
+R: http://www.r-project.org/ 
+Bioconductor: http://www.bioconductor.org/
+Tidyverse: https://blog.rstudio.org/2016/09/15/tidyverse-1-0-0/
+
+
+
+
+--------------------------------------------------------------------------------
+
+# Mail from Thomas, Sep 28th 
+ 
+ found 2 plates from previous courses. The data contains two channels:
+H2B, informative about chromosomes and tubulin, informative about the
+spindle.
+
+Segmentation and feature extraction in each of these channels allows us
+in principle to :
+- classify each "object" (chromosome or microtubule conformation,
+respectively) separately into one out of several predefined
+morphological classes
+- make a joint classifier by concatenating features from the two channels.
+
+As a result we obtain thus for each time point one or several
+classification results per cell.
+
+Now, in practice: I did not find classifiers for the two channels from
+the previous years. For this reason, I made a new classifier for these
+data on Friday and yesterday (so I annotated a number of cells), but as
+this takes a bit of time I only made one classifier in the H2B channel.
+
+I now run the analysis, as I understand that you would like to have
+representative data as soon as possible. We will see what we will do
+with the second channel afterwards, but at least you can start working
+on this.
+
+The analysis is now running, and I will send you results as soon as they
+are ready.
+
+Best,
+
+Thomas.
+
+
+# background on the experiments extracted from book on practicals from last years
+
+* ref: http://www.nature.com/nature/journal/v464/n7289/full/nature08869.html
+
+## Practical 3 will generate the data:
+
+In this practical we will learn how to set up automated time-lapse imaging of
+a multi-well plate. Specifically, we will image HeLa cells stably 
+expressing H2B-mcherry and tubulin-GFP, plated on 96-well plates that
+are coated with siRNAs causing mitotic phenotypes. 
+
+## siRNA scoring scheme
+
+To summarize the siRNA scoring: for each (TECHNICAL) replicate, we calculated one score in each
+morphological class (and joint classes) as the maximal difference over time between the time
+series in that class and the average negative control time series. The siRNA score was defined
+as the upper median of the corresponding replicate scores, and a gene was considered as a
+“potential hit” if at least one targeting siRNA had a score above the threshold. A gene was
+considered as a “validated hit” if two or more siRNAs resulted in a consistent phenotype.