DNA sequence input by voice

imsweetman at its.dundee.ac.uk imsweetman at its.dundee.ac.uk
Fri Jan 21 06:38:07 EST 1994


In article <2hmila$3ko at gazette.bcm.tmc.edu> pburch at cmb.bcm.tmc.edu (Paula Burch) writes:
>Path: dundee.ac.uk!doc.ic.ac.uk!agate!headwall.Stanford.EDU!bcm!cmb.bcm.tmc.edu!pburch
>From: pburch at cmb.bcm.tmc.edu (Paula Burch)
>Newsgroups: bionet.software
>Subject: DNA sequence input by voice
>Date: 20 Jan 1994 18:31:05 GMT
>Organization: Baylor College of Medicine, Houston, Tx
>Lines: 18
>Sender: pburch at roc.mbcr.bcm.tmc.edu (Paula Burch)
>Distribution: world
>Message-ID: <2hmila$3ko at gazette.bcm.tmc.edu>
>NNTP-Posting-Host: roc.mbcr.bcm.tmc.edu



>Now that microphones in computers have become commonplace, has anyone 
>developed a sequence input program that takes advantage of them?
>(Perhaps something for the Mac 840AV?)

>I think that a lot of people would vastly prefer reading a DNA sequence 
>into a file by voice than by any of the various clumsy and/or slow 
>and/or expensive devices currently in use.

>It was a disappointment to discover that the tantalizingly-named
>Digispeak works with a Grafbar! There are several programs that 'speak' 
>a sequence as it is typed in or entered by a Grafbar device--what we need 
>now is the opposite, something that will listen to a human voice.

>________________________________________________________________________
>Paula E. Burch, Ph.D.           Molecular Biology Computational Resource
>Baylor College of Medicine                  internet: pburch at bcm.tmc.edu   
>Houston, Texas 77030            phone: (713)798-6023  fax: (713)790-1275


Unfortunatly speech recognition is still an area which requires a lot of 
development before a *reliable* workable system will be available. As with 
most research in computers this might not take to long: but we're looking at a 
serious piece of AI type work and I would'nt hold my breath.

Before the shouts come back: this is'nt to say that some current speech 
recognition systems are not impressive: but, in my experience,  they are not 
reliable anough for an application as you describe: however: if your input is 
limited to ATUG (that may be nieve: I'm a systems developer not a biologist) 
that is a quite a *nice* limitation to impose on an experimental system... 
(brain now buzzing).

What, exaclty, would you require in such a system?

Iain Sweetman,
Dept of Anatomy & Physiology,
University of Dundee,
SCOTLAND.
imsweetman at dundee.ac.uk





More information about the Bio-soft mailing list