-
Notifications
You must be signed in to change notification settings - Fork 1
/
CHANGE.ME
123 lines (78 loc) · 4.47 KB
/
CHANGE.ME
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
CHANGE LIST, BUG LIST, AND WISH LIST FOR MAPMAKER VERSION 3.0 AND
MAPMAKER/QTL VERSION 1.1
CHANGES SINCE PREVIOUS RELEASE:
NOTE: Many copies of MAPMAKER 2.0 and MAPMAKER/QTL 1.0 were distributed as
Versions 1.9 and 0.9, respectively. They are basically the same, modulo a few
big fixes, and we refer to them collectively as "Release 2.0". All features
below pertain to MAPMAKER/EXP unless marked (QTL), in which case they apply to
MAPMAKER/QTL.
The license is changed to make it completely freely redistibutable.
Too many bugs were fixed to list here, and too many small features were cleaned
up to list here. This list is partial.
Command-line editing and history works on Unix (including A/UX). Try the up and
down arrows (or control-P, -N, -A, -B, -E, -K, -Y). The keystrokes are Emacs-
like, although we have not documented them here. It does NOT work in a Sun
cmdtool window, as described in the inkstallation instructions.
The "linked?" command is gone: Instead there is the "?" and "=" syntax in
sequences.
The "ripple" command lets you specify how many markers to permute.
Two-point data are always computed on demand: it is never necessary to execute
"two point" explicitly. This is NOT true of three-point, however.
Print names now works much better. Using marker names as arguments works much
better. However, marker names can no longer be abbreviated - this caused too
many problems.
All "genome analysis" features are new. This has had some effects on sequence
syntax. See manual.
All "error detection" features are new. See manual.
The old "three point" analysis features are gone, replaced with much better
ones, fully integrated with the multipoint analysis. See manual.
All "join haplotype" features are new. These are not mentioned in the turorial,
and are a first cut - they're a little rough around the edges yet. Use them
carefully.
All multipoint analysis is now based on the new "Hidden Markov Model"
algorithm, allowing new cross types. In some cases it's faster, in others
slower. Relative log-likelihoods should not have changed.
Postscript graphics added: "draw chromosome", "draw map", "draw scan" (QTL).
"Genotypes" command replaces "count crossovers". "count meioses" is gone,
replaced (sort of) by "suggest subset".
The "orders" and "build" (new) commands are now really useful for automatic
mapping. We have experimented with optimization approaches (e.g. Simulated
Annealing, etc), and we prefer the greedy algorithms.
The input file format has changed very slightly, if you either (i) used a
custom genotyping notation, or (2) placed map orders in your data file.
Modifying your raw files should take very little time. Only MAPMAKER (not QTL)
can prepare data now. See manual.
The "import data" and "export data" features are new, although they are
placeholders for those who which to build a database interface from/to
MAPMAKER.
The ".data", ".traits" etc. file formats have changed drastically. Release 2.0
files can NOT be loaded into the new versions.
CEPH analysis features are gone. Experimental crosses only for the time being.
Steve Lincoln has finally gone off to Grad School, and while he may do some
MAPMAKER development, Mark Daly will be the main front man.
KNOWN BUGS AND PROBLEMS IN THIS RELEASE:
Two-point analysis for F3 and RI is slow.
Control-C handling occasionally has problems corrupting internal data -- use it
sparingly. On DOS Control-C can sometimes abort MAPAKER, although this happens
more often under windows (WATCOM problem?).
Very large data sets cause sequence-parsing to work very slowly with print
names on.
The assign algorithm is correct and robust, although it is sensitive to global
state (which markers are already assigned) in a way that "group" is not. Not a
bug, but an issue.
TO-DOs and WISH LIST:
QTL needs many revisions. In particular, MAPMAKER/QTL needs the new MAPMAKER
converge_to_map() code, to understand other cross types, and QTL needs
qualitative models.
The Fast Three-Point feature did not make it into this release. We really want
this!
We need to refold outbred analysis into MAPMAKER, only better. QTL too.
MAPMAKER/QTL needs a model for polygenic qualitative traits.
A way to make command line editing work on DOS.
A way of importing small modifications to a data set
Draw editable Mac PICTs rather than PostScript
Native Mac Port
Better Windows front end?
Other Ports (Solaris 2.0, VMS, etc)
Integrate with Steve Knapp's, Piet Stamm's, and Ken Manley's software.
Ver 3b: S. Lincoln 12/92