Fragment Assembly Exercise

FRAGMENT ASSEMBLY LAB EXERCISE
Genomes & Genome Analysis
14 September 2000

Goal: To gain experience with the genome assembly problem and software solutions.

You will work with one of three teams. Each team will collaborate to assemble one or more contig(s) from the same fragments, but each will use a different assembly algorithm.

You will initially work with BAC sub-clones from Arabidopsis thaliana. You will not have to be concerned with vector screening, repeat masking, or per-base quality scores for these data. Contingent on availability of data from a bacterial whole-genome shotgun sequencing project, you may have an opportunity to assemble a whole genome!

Update: the data are in. You will assemble the entire genome of Chlamydophila pneumoniae AR39, which casues pneumonitis in mice. Thank you TIGR! We do not have repeat information about these data, so please just assemble them as you did the BAC data. It will take longer for the computer to run, because many more fragments and nucleotides (about 10x) are present. Work with the BAC data first until you are comfortable running the software and interpreting the output.

Please read carefully the documentation for your assembly algorithm to understand what parameters and input files it expects and what outputs it will produce. Interpreting the output is likely to be the most challenging task!

Each team is asked to co-author a written report that will be placed on the course web site to share with others. Please follow the form of Methods and Results sections in a scientific manuscript. See Read et al., cited below, for a published example. The due date is to be determined by the Professor. Students taking the 444 course are not required to co-author the report, but are encouraged to do so, if they choose.

The report is to describe the whole-genome assembly, if possible. You may want to address how the computational time required scaled with the increasing data volume.

In your reports, be sure to address the following:

What specifically did you do to perform an assembly?
What software did you use, from what origin, which version, and what parameters?
How did you choose those parameters?
Did alternative parameters (if any) yield comparable results?
How long did the algorithm take to run?
Summarize the outcome of the assembly.
How many contigs resulted, (how many gaps)?
What were the lengths of the contigs?
How did the assembly compare with the published result?
What method did you use to compare your assembly with that in Genbank?

NB: The purpose of the exercise is not simply to reproduce the Genbank version, but to know to what extent the outcome is affected by varied algorithms and parameters.

Warning: The data and software we are using for this lab were kindly provided under very strict licensing constraints. Please do not abuse this privilege. The data and source code are confidential, not to be redistributed or used without express conset of the copyright holders. Thank you.

About the Data

Arabidopsis thaliana BAC F16P2

located in the directory ~phraber/data/AtBAC
1 897 sub-cloned fragments
1 346 372 nucleotides
from chromosome II (2), section 238 of 255
GenBank accession AC004561