Markers, linkage groups and 2-points estimations

For a given dataset, CARTHAGENE performs all 2-points LOD calculations when the data set is loaded. This explains why loading large datasets can be (relatively) slow for expensive pedigree (eg. outbreds). All 2-points information can be requested using the mrklod2p, mrkfr2p and mrkdist2p commands. The mrkdouble (and mrkdoubleget which returns a Tcl list of compatible markers) command reports all pairs of markers that have compatible typing (i.e., markers which may be identical) along with their 2-points LOD. This can occur in two cases:

• the two markers have never been typed on a single common individual. These 2 markers should not be merged and this is clearly indicated by the fact that their 2-points LOD is equal to $0$.
• the two markers have been typed on a possibly large number of individuals and no sure recombination (or breakage for RH) exists between the two markers. The 2 markers can be merged and this is indicated by the usually strong 2-points LOD between the 2 markers.

When two markers are merged using the mrkmerge command, no information is lost. All the information is merged in the first marker and the second marker is deselected. The number of markers to map is lowered and the information stronger. This is always a good thing. Command mrkmerges merges all pairs of compatible markers detected by mrkdouble automatically.

Several other commands can be used to get information about markers. The most useful ones are probably the commands that can convert marker names to their numerical Id (mrkid and mrkids commands) and vice-versa (mrkname and mrknames commands).

The 2-points estimations are the basis of the linkage group computation process performed using the group command. This command puts in the same linkage group any two markers such that:

• their 2-points LOD is above a given threshold and
• their 2-points distance is below a given threshold.

The two conditions must be met together. Once groups are computed using this command, you can use a group using the groupget command.

If you have several datasets on the same markers (for instance, two RH panels with different resolutions), you can further eliminate markers which are poorly linked in at least one dataset to the rest of the markers in the selected linkage group. See the command multigroup.