Find-a-gene project
Introduction to Bioinformatics (260.602.01)
J. Pevsner
pevsner@kennedykrieger.org
Autumn 2006
The find-a-gene project is a required part of this course, and accounts for 30% of your grade. You should prepare a written report in a Word document that has the following components. The assignment is due the last day of class (Wednesday, October 25) as a hardcopy (not email). Steps [1] to [4] can be accomplished very quickly (at best within 5 or 10 minutes), so if you don’t succeed at first, just keep trying. I expect the work to be YOUR OWN, and the main point is for you to grasp the principles of database searching and sequence analysis that we cover in course. However, you CAN discuss the project with me, Hugh, Jen, Meg, or your classmates.
[1] Tell me the name of a protein you are interested in. Include the species and the accession number. If you do not have a favorite protein, select human RBP4.
[2] Perform a BLAST search against a DNA database, such as a database consisting of genomic DNA or ESTs. The BLAST server can be at NCBI or elsewhere. Include the output of that BLAST search in your document. If appropriate, change the font to Courier size 10 so that the results are displayed neatly. You can also screen capture a BLAST output (e.g. alt print screen on a PC). It is not necessary to print out all of the blast results if there are many pages.
On the BLAST results, clearly indicate a match that represents a protein sequence, encoded from some DNA sequence, that is homologous to your query protein. I need to be able to inspect the pairwise alignment you have selected, including the E value and score.
In general, step [2] is the most difficult for students because it requires you to have a “feel” for how to interpret BLAST results. You need to distinguish between a perfect match to your query (i.e. a sequence that is not “novel”), a near match (something that might be “novel”, depending on the results of step [4]), and a non-homologous result.
[3] Gather information about this “novel” protein. At a minimum, show me the protein sequence of the “novel” protein as displayed in your BLAST results from step [2]. In some cases, you will be able to do further BLAST searches to obtain even more sequence of your novel gene.
Here, tell me the name of the novel protein, and the species from which it derives. It is very unlikely (but still definitely possible) that you will find a novel gene from an organism such as S. cerevisiae, human or mouse, because those genomes have already been thoroughly annotated. It is more likely that you will discover a new gene in a genome that is currently being sequenced, such as bacteria or primates or protozoa.
[4] Prove that this gene, and its corresponding protein, are novel. For the purposes of this project, “novel” is defined as follows. Take the protein sequence (step [3]), and use it as a query in a blastp search of the nr database at NCBI.
--If there is a match with 100% amino acid identity to a protein in the database, from the same species, then your protein is NOT novel (even if the match is to a protein with a name such as “unknown”). Someone has already found and annotated this sequence, and assigned it an accession number.
--If there is a match with less than 100% identity, then it is likely that your protein is novel, and you have succeeded.
--If there is a match with 100% identity, but to a different species than the one you started with, then you have succeeded in finding a novel gene.
--If there are no database matches to the original query from step [1], this indicates that you have partially succeeded: yes, you may have found a new gene, but no, it is not actually homologous to the original query. You should probably start over.
[5] Generate a multiple sequence alignment with your novel protein, your original query protein, and a group of other members of this family. A typical number of proteins to use in a multiple sequence alignment is a minimum of 5 or 10 and a maximum 30, although the exact number is up to you.
[6] Create a phylogenetic tree, using either a parsimony or distance-based approach. Bootstrapping and tree rooting are optional. Use any program such as MEGA3, PAUP, or Phylip.
[7] Optional: compare the predicted structure of your protein to that of a known structure.
[8] Optional: show whether this gene is under positive or negative evolutionary selection.
[9] Optional: discuss the significance of your novel gene. What have you learned about this gene/protein family?