This preview shows pages 1–3. Sign up to view the full content.
This preview has intentionally blurred sections. Sign up to view the full version.View Full Document
Unformatted text preview: BCB/CprE/ComS 548 Fundamental Algorithms in Computational Biology Fall 2005 Homework 1 Solutions 1 - G T G C T G C- A G 10 5 10 5 10 5 G 10 5 15 10 5 10 5 C 5 5 10 25 20 15 20 A 5 20 20 15 15 T 10 5 15 30 25 20 The best local alignment, with a score of 30, is: G_GCAT GTGC_T 2 The point of this problem is show the impact of the scoring metrics on the alignment that is produced. The overall message is if you penalize it more, you will see it less. Some more specific observations are: With the default scoring, the gap opening penalty is very high, which results in very few gaps being introduced. Consequently, the only alignment that can be made is with the very short, but highly similar region, between the suffix of one sequence and the prefix of the other. Similarly, doing a local alignment with the default scoring metrics produces a very short alignment because gaps are so strongly discouraged. The second group of scoring values, with no gap opening penalty, allows for alignment between less similar regions. Also, since a mismatch counts the same as two gaps, there is no benefit to choosing a mismatch rather than just opening a gap in each sequence. The preference for inserting gaps also allows for a much longer local alignment, though not necessarily a very good alignment due to the number of gaps. The third set shows that increasing the gap opening penalty favors inserting one big gap over trying to align a region using a lot of small gaps. This creates the best local alignments. GRADING Any observations made were considered. In general, points given were based on the quantity and quality of your comments. I needed to see more than just increasing this penalty makes this statistic go up and that one go down. I wanted either an explanation of why it happens, or some conclusions about the effect on the significance of the alignment. 3 This is essentially a prefix-suffix alignment problem except that we do not want to allow any mismatches or gaps. This is easily accomplished by scoring mismatches and gaps as - and running the algorithm for the end-gap free global alignment: |s| = n, |t| = m...
View Full Document
- Fall '06