Fatal errors cause Mega2 to terminate without creating output files. We try to identify as many problems with the data files as we can before termination of a Mega2 run, for the user's benefit. However, several iterations may be necessary with wrong or corrupted data.
The most common problem users have faced so far has had to do with
DOS-format map files. DOS text files contain the "^M" carriage-return
character in addition to the "^J" line-feed character used by the Unix C
library routines. This caused the map file to be read in incorrectly by
earlier versions up to version 2.1 beta R3. This resulted in warning messages
such as :
WARNING: Locus GENE1 is not in map file.although GENE1 is actually present in the marker file. In extreme cases, Mega2 cannot find a single marker defined in the Map file and terminates abnormally.
So now each input file is first checked for "^M" characters, if these are present, Mega2 exits with an error message.
After the marker loci and traits are finalized, Mega2 performs consistency checks on the locus and pedigree files.
Mega2 checks to see if the locus file records have complete information, such as a name, required number of allele frequencies and whether these sum to 1.0 (a tolerance of 0.001 is allowed).
Mega2 reports problems such as incomplete pedigree records, invalid sex and proband fields, and missing parent records. Here are a list of error messages and their significance.
"Pedigree 100: entry ID (10) is undefined"
Check the pedigree file format, the pedigree records have been
corrupted.
"Pedigree 100: entry record 3 is corrupt or incomplete"
The record does not have all the necessary fields.
"Pedigree 100: father 1 of 3 is not male!"
"Pedigree 100: mother 2 of 3 is not female!"
This is flagged as a fatal error because in most cases, subsequent
analysis with the output data may not run.
"Pedigree 100: one parent of entry 3 is defined, the other is
undefined"
Most pedigree analysis programs require both parents to be defined,
therefore, so does Mega2.
"Pedigree 100: member 1 has unknown sex (0)"
1 and 2 are the only values allowed in the sex field, any other value
is flagged as an error.
Inconsistencies in genotype data, and one or more unconnected pedigrees with the same pedigree ID are flagged but considered non-fatal. The user is offered limited means of correcting the problem, or terminating Mega2.
Mendelian inconsistencies :
"Ped 100: Male 1 is a heterozygote at X-linked locus M1."
If the analysis is on Sex-linked data, and males have two different
alleles at a sex-linked marker, mega2 will flag this error. The user
can then choose to set the genotype to 0/0, or ignore this
discrepancy.
"Ped 100: Entry 1 has non-mendelian genotype at locus M1."
Mega2 reports Mendelian inheritance violations (for autosomal as well
as sex-linked loci). As before, the user can choose between keeping
these genotypes as they are or setting them to 0/0s.
Half-typed genotype fields :
"Pedigree 100: Entry 3 is half-typed at locus M1"
The person has one zero and one non-zero allele at that locus. User
may set both to 0/0, or keep them as they are, although it is always
better to stop and take care of this problem.
This menu is displayed if Mega2 found non-fatal problems with the pedigree data. It will display all three options or a subset depending on which problems were detected.
Found these errors in one or more pedigrees: Unconnected entries Mendelian-inconsistent genotypes Half-typed genotypes How shall we proceed? ========================================================== 0) Done with this menu - please proceed 1) Set mendelian-inconsistent genotypes to (0,0) [yes]. 2) Set half-typed genotypes to (0,0) [yes]. 3) EXIT Mega2. Select from options 0-3, 1 and 2 to toggle >
Some mendelian inconsistency checks may go undetected with Mega2, when there are sibships without typed parents. mega2 warns the user of this fact. In this case, the pedigree data should be analyzed with a more advanced pedigree checking program before running it through Mega2.