Events News Research CBS CBS Publications Bioinformatics
Staff Contact About Internal CBS CBS Other
CBS >> CBS Courses >> Brasil Workshop 2009 >> Course Programme >> Day 4

Day 3: Pan and core genome plots, Emboss and Blast atlases



Pan and core genome plots

Pan and core genome plots are graphs that display to what extent gene familes are conserved within a set of genomes. Conservation is evaluated by first BLASTing the proteomes of the genomes againt each other. This is done in a certain order, in that for every proteome, it performs a BLAST search against all previous proteomes. The result is a set of numbers specific for that time point that represents the proteome in the order of the input list, showing:

  • Number of new genes
  • Number of new families
  • Size of core genome
  • Size of pan genome
Two genes are considered to belong to the same gene family if the two are more than 50% identical over more than 50% of the length of the longest of the two genes.

We have prepared a script which produces such a pan- and coregenome plot, provided a list of proteomes.The result is a set of numbers specific for that time point that represents the proteome in the order of the input list. The script will accept a number of proteomes (pr1, pr2, .. prN) and perform a BLAST search of each proteome against all the previous:
  • pr2 against pr1
  • pr3 against pr1+pr2
  • pr4 against pr1+pr2+pr3
  • ...
  • prN against pr1+pr2+pr3 ... pr[N-1]
After these searches, the program will derive the number of core and pan proteins for each proteome. The output list will the be redirected into an R-script which plots all the core/pan values as a function of the proteome number. Just like the BLAST matrix script you tried yesterday, this script will cache all the BLAST results. In the event you change the order of the input proteins, all BLAST searches must be carried out again. Therefore, we have prepared two runs for you:

  1. First, log in and create a directory for this work.
    # log in to the computers again as, then:
    ssh -Y <your_username>@login.cbs.dtu.dk
    ssh -Y sbiology
    setenv MAKEFILES /home/people/pfh/bin/Makefile
  2. Create a directory where this work will be done.
    # Ensure you are in the right place
    cd ~/
    mkdir coregenome
    cd coregenome
  3. Create configuration file for this program
    # create config file
    sh ~karinl/scripts/core/coregenome.sh > pancoregenomelist.txt
    perl ~pfh/scripts/coregenome/coregenome pancoregenomelist.txt > pancoregenomeplot.ps
  4. Transfer file to your own computer and view it.

    Open a new Console window. In this window do the following:
    # Transfer file and view it.
    scp <yourUserName>@login.cbs.dtu.dk:coregenome/pancoregenomeplot.ps .
    gv pancoregenomeplot.ps
    Look at the plot. Can you tell how many gene families, approximately, your genomes have in common? How many gene families are there in total for your genomes?

EMBOSS

EMBOSS is a collection of software tools that are freely available at http://emboss.bioinformatics.nl/. Go and have a look – scroll through the the list of tools to get a feel what you can do with EMBOSS. If the site/analysis is too slow, ask me and I will tell you what to do.

  1. Using dottup from the EMBOSS package, see how colinear the three of your genome . Select two that you think are very similar, and another that should be very different from the other two.

    Find the refseq ID for your three genomes. Enter your ids in the box where it says To access a sequence from a database, enter the USA here:  
    The format of what you enter here is

    refseq:<refseq_id>

  2. Also set the window size to 18.

    Based on the results, which pair of genomes are more similar to each other? Does this conform with what you expected?

  3. What happens if you change a few of the dottup parameters (e.g. word size)?

  4. Using polydot from the EMBOSS package, see how colinear your genomes are (set word size = 20).

  5. Determine the following things using your genomes:

    The %GC content and the dinucleotide relative abundance using programs from the EMBOSS package for your three genomes.

    Based on the results, which pair of sequences is most similar?

    Also, check the “deltarho-website” at http://deltarho.amc.nl/cgi-bin/bin/index.cgi.

Blast atlases

Precalculated genome atlases

Configurable blast atlases