funded by 
ISCN Error Analysis |
 Deutsche
Version |
Errors in ISCN formulae are very prevalent: some 10% of the karyotypes in the Mitelman database contain errors. Most of them are due to insufficient statements. Very often, break points are denoted which do not exist.
Apart from such errors based on sloppiness or typos, there are errors caused by a lack of grasp. It is not very easy to understand the ISCN, it is often unsystematic, and more complicated rearrangements cannot be expressed with it (se also: ISCN Discussion).
Hence, during the analysis of karyotypes, the CyDAS software generates a series of error messages; if there are several errors in one karyotype, only the first error will be explained. We tried to render the error messages generally comprehensible, so that a skillful reader could grasp the type of error and how to correct it. This page is intended to offer further support to grasp the type of the error and to remove it..
Though that error might look like a typo, transposed digits, or something like that, we think it can be attributed to ignorance of the correct denomination of the chromosomal bands: in many cases, approaches to get to a correct band description by simple transpositions fail, e.g. 8q34: neither 8q43 nor 8p34 do exist. With 7q12, we could believe that 7q21 was meant, but of course it could be 7q11.2.
Often the author knows the breakpoints, but because they are generally
known he does not write them down. That's extremely common with the Philadelphia
translocation which is just described by t(9;22) instead of t(9;22)(q34;q11).
If the breakpoints were specified at another investigation of the same
patient, or somewhere else in the same publication, these breakpoints are
not repeated. For the software, the situation is clear: the breakpoints
are missing, the data are useless for further analysis.
The CyDAS desk top application offers the function "Edit
- Breakpoint Search" which automatically adds breakpoints for the cases
described above; it must be called explicitly. It is not available with
the internet applications.
Generally, the software searches for breakpoints in other clones of the same karyotype, i.e. if several clones are separated by slashes ("/") in the formula. During this serach for break points, errors may occur or erroneous aberrations may be created, thus giving raise to the latter error messages.
By the way, there are only two sites in the ISCN manual stating explicitly
that breakpoints need not be repeated:
An example is on p. 57 shwoing a karyotype with a Philadelphia translocation
and an extra Philadelphia chromosome: 47,XX,t(9;22)(q34;q11),+der(22)t(9;22)
"The breakpoints in the extra der(22) need not be repeated." I.e.
the rearrangement giving raise to the derivative chromosome was denoted
in the same clone, and the type of rearrangement including the affected
chromosomes was still repeated.
The other site is an example on clonal evolution on p. 80 in the context
of the "idem" symbol.
With a karyotype 47,XX,t(9;22)(q34;q11),+der(22) it may be clear to the author that the "der(22)" means "der(22)t(9;22)(q34;q11)". But how is a software expected to guess that? It could also mean "der(22)add(22)(p11)del(22)(p11)".
With polyclonal karyotypes, the CyDAS software tries to find the aberrations from other clones. With 47,XX,t(9;22)(q34;q11),+der(22)t(9;22)(q34;q11)/48,XX,+8,t(9;22)(q34;q11),+der(22), such an expansion works. With 45,XX,-9,der(22)t(9;22)(q34;q11),der(22)t(9;22)add(22)(p11)/46,XX,+8,-9,der(22), a surely correct expansion is no more possible. During such expansions errors may occur or erroneous aberrations may be generated, thus giving raise to the latter error messages.
Even if the fragment involved is very small only and does not extend over the borders of a chromosomal band, two breakpoints are required! A duplication of Xq21 is written dup(X)(q21q21) but not dup(X)(q21).
By the way, there is quite some difference between a del(5)(q13) and a del(5)(q13q13): the former describes a terminal deletion of the q arm from band 5q13 upto the q terminus, the latter one describes an interstital deletion of a tiny fragment residing inside band 5q13 but keeping all the rest. Since both notations per se are correct - but with such extremely different meaning! - the CyDAS software cannot generate a useful error message here.
When scrutinizing some examples from the Mitelman database, we can deduce that a "+t" should mean the gain of a dicentric chromosome, i.e. a "+dic" is meant. Especially from the given chromosome count can be deduced that only one chromosome was gained.
E.g.
Chadduck
et al 1991, Pediatr Neurosurg, Case 1
47,X,-Y,-2,+t(2;12)(q13;p11),t(4;17)(q31;q25),-12,+3mar
could be
47,X,-Y,-2,+dic(2;12)(q13;p11),t(4;17)(q31;q25),-12,+3mar
or better
47,X,-Y,dic(2;12)(q13;p11),t(4;17)(q31;q25),+3mar
since the dicentric chromosome replaces a chromosome 2 and 12 each.
Similarly, a gain of an insertion or a telomeric association does not make sense.
There are also translocations involving three chromosomes ("three way
translocations"). An example from the ISCN manual is
46,XX,t(2;7;5)(p21;q22;q23)
Here, the f ragment 2p21pter moved to band 7q22 and
replcaes all the chromosomal material from there upto theq terminus of
chromosome 7. That region
7q22qter in turn moved to band
5q23
and replaces everything from there upto the q terminus of chromosome 5.
The latter region 5q23qter finally moved to band 2p21
and replaces the fragment 2p21pter.
Very often, descriptions are encountered with two bands (i.e. a frgament)
for one chromosome of the three chromosomes (and one band each for the
other two chromosomes). E.g.
Jin
et al 1995, Cancer Res, Case 43
46,XY,t(7;10;15)(q11;p11q26;p11)
It cannot be a normal three way translocation because then there was
only one band for chromosome 10 (i.e. either 10p11 or
10q26).
Nor can it be a translocation of interstital fragments because then also
the chromosomes 7 and
15 required two bands each.
From communication with other cytogeneticists, we deduce that some people
want to state that the chromosome for which two bands were given was involved
in two translocations which were written as one term. That means for the
above example that there occured the translocations t(7;10)(q11;p11)
and
t(10;15)(q26;p11) which both involved the same chromosome
10.
Hence, the derivative chromosome 10 suffered two aberrations,
and ought to be denoted correctly as: der(10)t(7;10)(q11;p11)t(10;15)(q26;p11).
Accordingly, both the other involved chromosomes 7 and
15
must be denoted as derivative chromosomes:
der(7)t(7;10)(q11;p11),
der(15)t(10;15)(q26;p11).
The corrected formula then is:
46,XY,der(7)t(7;10)(q11;p11),der(10)t(7;10)(q11;p11)t(10;15)(q26;p11),der(15)t(10;15)(q26;p11)
But if there are two distinct translocations which both involve one common chromosome, all three involved chromosomes must be written as derivative chromosomes with the aberrations involved in their respective genereation.
If the chromosomal break occured not exactly in the centromere but close by it, the next band has to be denoted, i.e. p11 or q11 (or more excatly, e.g. p11.1). Here, we must consider if the centromere was part of the translocated fragment thus generating a dicentric chromosome which must be dentoed with a "dic" symbol, or not.
A translocation of a whole chromosomal arm exactly from its centromere to a position on an other chromosome outside the centromere does not make sense: either the break was just behind the centromere generating a chromosome which contains the centromere of the other translocation partner only, or the break was just before the centromere thus translocating a small portion of the other arm also and generating a dicentric chromosome (such translocations are always non-balanced because an acentric fragment remains which will be lost during the following cell divisions).
Examples:
These are mainly two types of errors:
A non-existing symbol was used for the description of an aberration.
The cause might be a typo, e.g. writting ibs instead of ins.
In some case, the search pattern already requires a specific breakpoint
description, e.g. for isomerisation. An isomerization can only occur at
a centromere, i.e. the band must described as p10 or q10.
If p11 was named instead (iso-dimerisation), the isomerisation
could not be discovered.
Sometimes, wrong symbols or notations can be intercepted at an earlier stage: if they contain characters which are not part of any symbol. Here, the latter message is issued.
If some characters remain after the analysis of an aberration, that
aberration is criticised as invalid. Often just a comma is missing which
separates this aberration from the next aberration in the ISCN formula;
then these two aberrations are regarded as one aberration which could no
more be analysed correctly.
Warning: with a derivative chromosome being the first aberration, such
a composite term could be analysed as a correctly described term!
But often, there is a general lack of knowledge of the ISCN, as can be proved with the CyDAS log files:
Many cytogeneticists believe that the detailled ISCN notation is no more valid and then try to describe a highly aberrant chromosome using the short notation (that opinion is definitley wrong!). Often it is sufficient to write down all the aberrations which generated the derivative chromosome after the "der" symbol; but a number of cases remains when that short notation fails (see ISCN Discussion).
More often, the overview seems to get lost with such chromosomes, and hence some bands are assigned wrongly.
The CyDAS software calculates a derivative chromosome step by step, introducing the aberrations one at a time in the order given. Often the software discovers that a band which is required for a rearrangement (a breakpoint) is not present in the chromosome.
Examples:
Because such notations are very common, we must assume that is not (always) a typo which mutated a deletion ("del") into a derivative chromsome.
More likely, the authors wanted to tell that they identified a chromosome which has some deviation, and that they also could narrow down the site where the deviation is located.
Maybe the authors wanted to express that chromosomal material of unknown origin replaces the material from that sit upto the terminus of the same chromsome arm. That is the hallmark of an addition which is to be described with the symbol "add". E.g. add(14)(q?) instead of der(14)(q?).
It is also possible that chromosomal material of unknown origin is located at that position, followed thereafter with the normal rest of the chromosome arm. That is called an insertion of unknown material, which is described with the symbol "ins". E.g. ins(3;?)(p?;?) instead of der(3)(p?).
The situation is more complicated if the origin of the derivative chromosome proper is not clear (i.e. if we cannot determine the origin of teh centromere), but when we can determine that somewhere material of the chromosome denoted follows, and maybe thereafter again unknown material. Such an aberration is a gain of a derivative chromosome of unknown origin with a translocation or an insertion. E.g. der(?)t(7;?)(q?;?) stands for a der(?)(?->?cen->?::7q?->7qter), or der(?)ins(?;14)(?;q?q?) stands for a der(?)(?->?cen->?::14q?->14q?::?).