Misuse of "homology" or how can you be half-pregnant ?

Robert Herzog rherzog at dbm.ulb.ac.be
Mon Nov 12 05:06:18 EST 2001


Hi,
A look at the recent announcement of the Fasta-blast scan software in 
this list prompts me at writing what follows.
In the very centre of the screen shot of the Fasta-blast scan program 
sits an alet box asking you about the "percentage of homology" you are 
looking for...
The term "homology" refers to the status of any object as being similar, 
or occupying a similar position in a bilateral comparison. Take a look 
at your Webster's or other dictionary for a comprehensive definition.
In no case can two items, objects, persons or whatever be "homologous" 
to a certain extent, which could be evaluated in percentage, or whatever 
else measure.
The term "homology" is heavily misused in modern bioinformatics, in the 
place of accurate terms "similarity", "identity" or others. Homology 
should be reserved to "corresponding organs, parts of organs" (from 
Webster's), and by extension, proteins, functions or genes  "determined 
by agreement in derivation and development from a like primitive origin" 
(same source...).
Human and rat hemoglobins are undoubtly homologous, and by extension, 
their sequences can be said to be homolous, and these sequences present 
a certain level of similarity or identity that can be expressed in 
percentage.
On the other hand, Mao Tse Toung and Queen Elizabeth of England were 
"homologous" (even if they might have been shocked by this 
statement...), as they were at one time heads of two countries and 
occupied as such equivalent (homologous) positions.

In the world of bioinformatics, let us reserve the term "homology" to 
situations where two bio-objects (be it sequences, genes or whatever) 
can be claimed to be derived from some common ancestor, or at least can 
be inferred to have the same function in their corresponding organisms.

Two women could be similar to a certain extent, like one being 15 
percent taller that the other one, but there is no way a woman could be 
40 percent pregnant, or more pregnant than another one... She is, or she 
is not in that condition-full stop. If she is, she is homologous to all 
other pregnant women on earth.

I hope my broken English was precise enough to pass the message. But 
this should not be a reason for you not to have a look at the 
Fasta-Blast Scan software...;-)
Robert Herzog
-- 
==========================================
Robert Herzog   -  (rherzog at ulb.ac.be)
Bioinformatics Laboratory, Dept.Mol.Biol.
Universite libre de Bruxelles, CP300
Rue des Professeurs Jeener et Brachet, 12
B 6041 Gosselies - Belgium
Tel : 32-2-6509978   Fax : 32-2-6509998
==========================================

---





More information about the Bio-soft mailing list