Bioinformatics and discovery:
induction beckons again
John F. Allen
With the flood of information from genomics, proteomics,
and microarrays, what we really need now is the computer
software to tell us what it all means. Or do we?
Introduction
In the life sciences, there has recently been a strong
resurgence of the view that there is a direct route from
observation to understanding. By this route, knowledge can
flow securely from data without the human and fallible
intervention of guesswork, imagination or hypothesis. Infor-
mation technology now puts oceans of data at our immediate
disposal, and even the ubiquitous personal computer can
process and analyse these data at huge speed. Surely, the
thinking goes, we can now expect computer programs to
derive significance, relevance and meaning from chunks of
information, be they nucleotide sequences or gene expression
profiles. A Nature editorial,
(1)
for instance, discusses biolo-
gists' increasing reliance on computers to do their thinking for
them. The editorial is rather kind to the biologists. Its titleÐ
``Can biological phenomena be understood by humans?''Ð
provocatively implies that scientific discovery might well be
carried out, instead, by machine. In contrast with this view,
many are convinced that no purely logical process can turn
observation into understanding. We owe this conviction, first
and foremost, to the work of Karl Popper.
(2±4)
Here I argue that
Popper was correct, and outline the way in which I think his
philosophy applies to bioinformatics. I predict that even the
formidable combination of computing power with ease of
access to data cannot a produce a qualitative shift in the way
that we do science: the making of hypotheses remains an
indispensable component in the growth of knowledge.
The problem of induction
``Logical deduction'' is a process by which the truth of a general
statement entails the truth of a particular statement. For
example, if it is true that ``all men are mortal'', then we can
deduce from the statement ``Socrates is a man'' that ``Socrates
is mortal''. The reverse process, a logical route from the
particular to the general, has been called ``logical induction'',
but it has never been clear how this might work. The possibility
of logical induction was dismissed by the Scottish philosopher
David Hume, in the eighteenth century.
(5)
One of Hume's
concerns was the idea of causality Ð how can we know that
``a'' causes ``b'', when all we can say with certainty is that we
have observed that ``b'' follows ``a'' on a number of occasions?
How many times do we have to observe that ``b'' follows ``a'' in
order for us to be sure that ``a'' causes ``b''? Hume's answer is
that we never can be sure. And what are we doing when we
make predictions about future events? For example, why do
we believe that the sun will rise tomorrow? Admittedly, we have
seen it rise many times before, but extrapolation is always
uncertain, and we feel that ``knowledge'' must be more secure
than this. Hume believed that we can never really know that the
sun will rise tomorrow. Our expectation that it will, like our idea
of causality, has, according to Hume, no rational foundation.
Bertrand Russell put the consequences thus: ``It is im-
portant to discover whether there is any answer to Hume. If
not, it follows that there is no intellectual distinction between
sanity and insanity. The lunatic who believes that he is a
poached egg is to be condemned solely on the ground that he
is in a minority, ... or on the ground that the government does
not agree with him''.
(6)
Russell also pointed to the stark
consequences of having no rational basis for the resolution of
conflicting theories. Writing in 1944, Russell put it thus: ``The
growth of unreason throughout the nineteenth century and
what has passed of the twentieth is a natural sequel to Hume's
destruction of empiricism''.
(6)
Induction and verifiability
In the early twentieth century, logical positivists proposed that
there was an answer to Hume, and that there was indeed a
logical route to certain knowledge. This route was ``scientific
method''. Science, and science alone, could tell us whether ``a''
causes ``b'', and allow us to predict when the sun will rise.
According to the philosophy of logical positivism, a general
statement or theory can be arrived at by inductive reasoning.
Positivists also thought that such a theory, if it is verified by
observation or experiment, can be promoted to a ``law''.
Indeed, positivists required that a theory must be verifiable in
order to count as ``scientific''. Verifiability was the criterion of
what is, and is not, science. Thus, in the view of positivists,
104 BioEssays 23.1 BioEssays 23:104±107, ß 2001 John Wiley & Sons, Inc.
Plant Biochemistry, Lund University, Box 117, SE-221 00 Lund,
Sweden. E-mail: john.allen@plantbio.lu.se
Funding agencies: Crafoord Foundation; Swedish Natural Sciences
Research Foundation.
Commentary