heapequiset

Activates or desactivates the detection of ``equivalent'' orders on datasets merged using the dsmergor command.

Synopsis:

The heapequiset command is invoked either as:

• heapequiset Options
• heapequiset on/off flag

Description:

When two datasets are merged, it is often the case that the two data sets only share few markers. When the two such datasets are merged using dsmergor, many different orders correspond to equivalent situations on the two datasets and they will get the same log-likelihood. Imagine for example that the dataset 1 is informative on markers $A$, $B$, $D$ and $E$ and that the dataset 2 is informative on markers $A$, $C$ and $E$. Then the orders $ABCDE$, $ACBDE$ and $ABDCE$ are equivalent and will get the same loglikelihhod since they both correspond to the two same orders: $ABDE$ on dataset one and $ACE$ on the dataset 2.

This phenomenon will populate the heap with many equivalent maps and also severely limit the efficacy of the buildfw process we uses the heap structure internally to detect if two orders with close enough likelihoods exist. At saome point, It may be impossible to insert a marker in the framework map because equivalent orders (therefore with the same likelihood) exist in the heap. To avoid this, activate the heapequiset command. It is important to mention that as soon as the mode is activated, ``equivalent'' orders are expunged from the heap. The command is still in alpha stage.

Arguments:

• Options : -u to obtain the synopsys of the normal use, -h to print a one line description, -H to print a short help.
• level: the on/off flag (0 for off or 1 for on).

Returns:

nothing.

Example:

# we load and merge two datasets
CG> dsload Data/bc.cg
...
CG> dsload Data/rh.cg
...
CG> dsmergor 1 2
...
# we evaluate two different but equivalent orders
CG> mrkselset {17 29 30 18 39 19}

CG> sem

Map -1 : log10-likelihood =  -164.94
-------:
 Set : Marker List ...
   1 : MS17        MS18     MS19
   2 : MS17 G9 G10      G19 MS19

# we move the marker 18 which is only defined in the first dataset
CG> mrkselset {17 29 18 30  39 19}

CG> sem

Map -1 : log10-likelihood =  -164.94
-------:
 Set : Marker List ...
   1 : MS17    MS18         MS19
   2 : MS17 G9      G10 G19 MS19

# the two orders have entered the heap
CG> heaprint

Map  0 : log10-likelihood =  -164.94
-------:
 Set : Marker List ...
   1 : MS17        MS18     MS19
   2 : MS17 G9 G10      G19 MS19

Map  1 : log10-likelihood =  -164.94
-------:
 Set : Marker List ...
   1 : MS17    MS18         MS19
   2 : MS17 G9      G10 G19 MS19

# we now activate the detection of equivalent orders
CG> heapequiset 1

# the equivalent order is expunged
CG> heaprint

Map  0 : log10-likelihood =  -164.94
-------:
 Set : Marker List ...
   1 : MS17        MS18     MS19
   2 : MS17 G9 G10      G19 MS19

# we evaluate a third equivalent order
CG> mrkselset {17 18 29 30  39 19}

CG> sem

Map -1 : log10-likelihood =  -164.94
-------:
 Set : Marker List ...
   1 : MS17 MS18            MS19
   2 : MS17      G9 G10 G19 MS19

# it did not entered the heap
CG> heaprint

Map  0 : log10-likelihood =  -164.94
-------:
 Set : Marker List ...
   1 : MS17        MS18     MS19
   2 : MS17 G9 G10      G19 MS19

CG>

See also:

• dsmergor (2.2.3)