Exercise 6 - Compare Gene Prediction Programs

To predict genes using gene prediction software it is important to understand the strengths and limitations of the different programs. This exercise will help you to evaluate how three different software tools perform.

• This sequence from hs18q (4,608 nucleotides) requires computational analysis. To gain a better appreciation for the relative performance of the methods used in gene prediction and how the results may vary between methods, use FGENESH, GENSCAN, and HMMgene (links under above) to answer each of the following questions:

  
  

  1. How many exons are in this sequence?

  2. What are the start and stop points for each of these exons?

  3. Which strand (forward or reverse) are the putative exons found on?

  4. Are there any unique features present, like poly-A signals or tracts? Where are they located?

  5. Can any protein translations be derived from the sequence? What is the length (in amino acids) of these translations?

  6. For HMMgene only, can alternative translations be computed for this particular DNA sequence? If so, give the number of exons and the length of the coding region (CDS) for each possible alternative prediction. Note on which strand the alternative translations are found.

  During your work make sure that you save your results in a word processor, as entire web pages, and/or as screen shots.

• Compare the results of the different tools.