Sequence comparisons (Bioinformatics?)

In the thread "Is convergent evolution evidence against common descent?" the topic of comparing genetic and protein sequences came up:(Message 138)This is, of course, relevent to that topic but it turns out there is enough information and resources on the web to have some fun with this. This topic is to spin off this area alone from the convergent topic.The ClustalW form at http://align.genome.jp/ , allows for comparisons to be made.I would like this topic to contain explanations (in considerable detail) of what is done by the program at that site and more on how to use it.We can also offer up questions about and examples of sequence comparisons here.

I have a rough idea of what is going on with the CLUSTALW program but I would like a 3 or 4 paragraphs explanation if WK (or someone could give that). The help is obviously aimed only at those "in the business" and is heavily jargon loaded.I presume that the program attempts to "line up" the input sequences as best as it can. It then calculates a percentage of match. (The PIM "percentage identity matrix").Correct?I would like some hint of what it means, how it may be correctly used and how it may be incorrectly used. What anomolies might one create?Why is the percentage called a percentage identity matrix?

What are those respects? What does having similar percentages mean? What is "too close"? What is not close enough?Are these not all relative numbers to position animals relative to each other?There are some proteins that are identical across a wide range of animals are there not? If we picked those would a series of 100's result? Would that matter or be meaningful? In what way?(I'm not called NOSY for nuttin' you know. )

In what way are all these used to determine a tree? Is there somewhere explaining it in some detail?Thanks.