Interval_Mapping

Menu items:

• Quit: quits the window.

• Help: brings up this help.

• Print: prepares a Post Script file in ./PS which you can then print on your local laser printer.

• Parse database: Must be run once, when you have changed the 2_point_data It looks for deficiency data and reorgaizes them internally for later use by this package. In particular, genes and def are sorted into contigs defined as a set of genes and deficiencies tested against each other.

• Read file: This allows you to run the same analysis on some data that you do not want to parse into your database. The data file should consist of ND+1 lines of NM symbols 1 or 0 where NM is the number of markers and ND the number of deficiencies. A 1 indicates that the deficiency uncovers the gene. The first line indicates which genes to take into account. It should have NM 1 to use everything, any 0 will discard the corresponding column. i.e. gene from the analysis. One can also impose the order of subgroups and block some genes together. An example input file is given in the next help page.

• Get contig: Ask for a contig number. If you give a bad answer, it does not matter. The number of contigs is specified when you Parse the database.

• Cost: Gives the cost of the running order of the marker, The cost function is n for a deficiency splitted in n+1 discontinuous pieces There is also a way to impose the order of some of the markers, but I this is not yet done in a smooth way.

• Shuffle: Reshuffles the markers before a new try.

• Monte Carlo: Inserts the markers one by one in an expanding list. It tries to keep the cost to a minimum and remembers n previous configurations, n is a parameter asked from the user. We had fast and good results with NM = ND = about 20 and n = 5 This is a monte carlo, so you must repeat the search several times. Once a good solution is found, it will no longer print the bad ones. With NM = ND around 100, n = 8, it was very slow but the results were rather good.

• Intrinsic tree: This is an independant fast and innacurate algorithm to reorder the genes. It should be completed by the next button:

• Refine tree: This should better the results of the preceeding button but is not yet available.

• Map: Creates a new window with a useful graphic presentation of the data. It is interesting even on the standard data, before selecting the order option. The Map graph has its own menu with the usual options Quit-Help-Print. The Help is the same one. Quit quits the application. Print prepares defmapxxx.PS. With NM = ND = 100, the resulting graph will not be printed on my laser, probably it is a font problem, sorry.

• Histo: Plots the marginal ditributions : Histogram of the number of deficiencies carrying a given number of markers and vice versa

last edited: July 1994