SAGE4 parameter file was wrong. The field-widths were not set correctly, caused the pedigree data to be read in incorrectly.
X-linked analysis was not working, caused Mega2 to crash.
Install script used bash.
SOLAR bug fixed: In the "loop-over traits" mode person ids in the phenotype file did not match person ids in the pedigree file.
SOLAR bug fixed: Mega2 crashed while looping over quantitative loci.
RECODE bug fixed: Mega2 crashed while recoding loci from names file, if there were no mapped loci.
MERLIN bug fixed: Merlin analysis option menu was displayed twice.
GENOTYPING ERROR simulation mode bug fixed: Locus selection was not working properly.
SIMWALK2 bug fixed: Location scores option produced an incorrect penetrance file for multiple affection liability classes.
ASPEX bug fixed - nuclear pedigree naming/numbering options did not allow numbering with extensions.
ASPEX bug fixed - nuclear pedigree naming/numbering options did not allow numbering with extensions.
SPLINK bug fixed - did not put all markers to be put into one file, when asked to do so.
S.A.G.E 4.x bug fixed - The format statement in the parameter file was incorrect.
R-graphics bug fixed - The output file was not renumbered according to the chromosome correctly.
S.A.G.E. 3.x affected relative counts files are not set up correctly.
S.A.G.E. 4.x parameter file did not have the correct pedigree field formats.
PREST pedigree file was not set up correctly. It contained one big line.
Merlin/SimWalk2-NPL option did not work due to a mismatch in pedigree naming between the two sets of files.
Pedigree and individual id menu behaviour has been fixed. The different options were not being selected and displayed correctly.
Fixed SPLINK format bug - was writing individual ids in place of mother's ids.
Fixed bug in SimWalk2 for handling more than 100 liability classes.
Fixed bug in the Win 32 version of Mega2, where the locus order line was not read in correctly. The 2.5 version will work correctly however, al lomg as there is an extra non-numeric character before the newline.
Blank lines are allowed before the locus-order line again, the check for locus order in Ver 2.5 did not accept blank lines just before the locus order line.
In SOLAR, the phenotype file and marker genotype file don't match the person ids in the pedigree file (id choice was not implemented for the former).
In locus reordering, for multiple chromosomes and multiple traits, traits were being output at the beginning of the order irrespective of the user's ordering.
Selection of loci for random genotype error detection from the batch file was incorrectly implemented.
> Renaming chromosome number extension for some output files was wrong.
Segmentation fault for 0 recombination fractions (where there is only a single locus in premakeped-format output files.
Existing vstrm files were not removed before each Vitesse run. This would cause the old file to be used when Vitesse failed to create a new one.
SimWalk2 penetrance file had invalid liability class values. Acceptable values are 1-99.
SPLINK option produced garbled pedigree files if the user selected the option to combine all markers into one pedigree file.
For SOLAR, the HHID phenotype was being written into the phenotype file, this has been moved to the pedigree file.
Bug in loop reconnection - Mega2 assumed that the proband numbers were ordered in the pedigree file which caused the numbering of links (first_off, sibs etc.) to be wrong.
Individual-count for pre-makeped format input pedigrees (in the screen output as well as log file) was wrong.
Selecting only trait loci in S.A.G.E. 4.0 resulted in a segfault.
Versions 2.3 and 2.3 R1 may have had a segmentation fault when creating nuclear pedigrees from post-makeped pedigrees which did not have person ids numbered consecutively from 1 thorugh the number of pedigree members. Nuclear pedigrees are also created in the S.A.G.E. 3.0, ASPEX, and GHMLB options. Pre-makeped pedigrees do not cause segmentation faults in either version 2.3.
Post-makeped pedigrees with non-ordinal entry ids may have also created segmentation faults in the "create summary" option for segregation summary. Again pre-makeped files are not a problem.
Mendel pedigree file format statement has been fixed. The format statement for reading in pedigree members was incorrect.
The script sage_cnt.*.sh was not working in case of a single affection trait. This script has been corrected.
In some cases, nuclear families included half-sibs This bug affects Aspex, Affected Sib Counts, GHMLB and S.A.G.E.
Loop reconnection had a bug. If the individual with the proband field set to 1 appeared AFTER individuals with other proband numbers, the loops were not reconnected.
Computation of sample variance for QTLs was biased before.
HWE option crashed if any marker locus had less than 3 alleles using GUO's hwe program.
Spurious inheritance errors for X-linked loci namely too many distinct alleles, even when there were at most three alleles in a mixed-gender sibship.
S.A.G.E. 3.0 fsp and fcor file formats were wrong. Shell file alos had a bug. Handling of multiple trait loci was not correctly implemented.
The LOG and ERROR files were appended to in each run, instead of being overwritten.
Segmentation fault if no loci were selected.
The Mendelian inconsistency checking routine produces spurious inconsistency messages because it does not handling half-siblings properly.
The "reset genotypes" menu does not EXIT, but continues to loop. After a few turns this will result in a buffer overflow as it keeps appending to the message at the bottom of the menu.
This is a list of bugs reported by users of Mega2 2.1 beta:
No spacing between the liability status and liability class
if the selected affection locus had multiple liability classes. Affected
options are :
4) Convert to GeneHunter-Plus format
5) Convert to GeneHunter-format
9) Convert to SLINK-format
11) Set up for homogeneity analyses
12) Convert to SIMULATE-format
17) Convert to Vitesse-format
18) Convert to Linkage-format
20) Convert to Allegro-format
21) Convert to MLBQTL format
The sib-phase XMGR scripts do not work for a grid of risk-ratios, the curve-data is missing the second column.
The Vitesse LOG and STM files may contain bogus locus names and numbers, and, therefore, may not produce correct results.
The fix to the previous bug made locus numbers ordinal (i.e. 1..N, where N is the window size, including trait) with trait locus always as locus number 1. This prevents the program "map" from creating the correct combined map from all these sliding windows. So, the locus numbers have been changed back to the original numbers. Version beta 2.1 has an additional bug in this ordering, so use Version R3 for a complete fix.
A space was missing between the affection status column and the next locus data column in the SPLINK output pedigree file.
Mega2 ran into infinite loops trying to sort pedigree members if the structure was highly complicated with inbreeding and matings across generations.
SLINK option crashed if the only trait locus chosen was a QTL.
Mendel output the first marker locus twice in the locus file if no afection locus was selected.
Allele count option did not skip untyped markers, so it gave meaningless output due to divide-by-zero errors.
Aspex and Splink options did not have the correct headers in the c-shell scripts when more than one analysis was put into the same file.
Last updated : June 15, 2001