Upload
haim
View
212
Download
0
Embed Size (px)
Citation preview
Massachusetts Institute of Technology.“Thereis tremendous enthusiasm from all quartersto finish the mouse genome, and scientistswon’t have to wait very long,” he says.
But he admits that “the importance of a lotof different projects will have to be juggled,and it’ll all get done in an order that ismaximally useful to scientists as a whole”. Likethe NIH, Germany and France are waiting toassess the scientific promise of differentspecies before committing their nationalsequencing efforts to finishing the mouse.
Jordan points out that the publicconsortium, concurrent with providing aworking draft, will finish selected areas ofthe mouse genome judged to be biologicallyimportant. Although primarily intended tosupport priority areas of mouse genetics,this “will also give us a better idea of howmuch better it really [would be] to have thefinished sequence”, she says.
Mouse geneticists need no suchconvincing. “A draft sequence will be a poortool for mouse geneticists,” says Phil Avner,head of the mouse molecular genetics unit atthe Pasteur Institute in Paris. It will allow nomore than 80 per cent of genes to beidentified, and will not provide keyinformation about areas of the genome thatregulate the genes, he says.
Even some zebrafish geneticists agreethat finishing the mouse genome is a highpriority because of the uniquely advancedlevel of genetic tools available. “Butzebrafish biologists will also need a workingdraft within the next year or two,” saysWolfgang Driever, a zebrafish geneticist atthe University of Freiburg. “So ultimately itwill be a question of timing and funding.” ■
Paul Smaglik,Washington Data on the human brain will soon be avail-able over the Internet via an electronic gate-way. Currently being planned by aninternational consortium, the portal willgive researchers access to data at variouslevels of detail and sophistication.
Backed in principle by the Organisationfor Economic Cooperation and Develop-ment (OECD),the consortium aims to devisecommon standards and procedures for thevarious neuroinformatics databases and soft-ware programs scattered around the world.
An OECD working group set up to dev-elop the gateway met for the first time inGenoa last month, and announced that itplans to make decisions on a launch date at itsnext meeting in Washington in September.
“We’d like to create a portal that wouldprovide access to all the resources that are outthere,” says Stephen Koslow, director of theOffice on Neuroinformatics with the USNational Institute of Mental Health, andchair of the OECD working group.
Koslow points out that many individualprograms already exist, each working at theirown levels of analysis. Finding them andlinking them to one site will be the first step,he explains. The next step will be to makethem work with each other.
The project resembles earlier efforts tobuilt computational tools for studyinggenomes.Many groups independently creat-
news
NATURE | VOL 405 | 8 JUNE 2000 | www.nature.com 603
ed computer programs that eventuallybecame consolidated into software suites.
But neuroinformatics covers a broaderspectrum of tools than computationalgenomics,ranging from brain imaging to thestudy of genes and proteins. And tyingtogether data dealing with different levels ofanalysis is a huge challenge. Data can be inmany forms, including brain scans showingthe development of Alzheimer’s disease andgenetic databases detailing susceptibility tothe disease.
Scientists involved in the Human BrainProject — a US government-backed effort todevelop a variety of neuroinformatics tools— are already working to make the variouspieces of software more accessible.
Jonathan Cohen, professor of psycholo-gy at Princeton University,said last week thatalthough independence in the early days ofneuroinformatics software developmenthad resulted in many useful programs, it hadalso created an electronic ‘Tower of Babel’.
“Since a lot of people build these thingson their own, they are not always in a formatthat everyone can use,” Cohen told a confer-ence on the Human Brain Project held lastweek in Bethesda,Maryland.
Cohen’s group is providing ‘wrappers’ tomake different programs look the same on acomputer screen. So far, this ‘FisWidgets’project has provided interfaces for 43 public-domain neuroimaging programs.
Another problem is that data can be in dif-ferent forms, or images in different resolu-tions.This is especially daunting when devel-oping multiple-scale models of the nervoussystem,Nigel Goddard,a bioinformatician atthe University of Edinburgh, told the meet-ing. Such modelling provides a broad map ofa system, along with the ability to zoom in onselected parts for greater detail, and retrievedata associated with that image.“We need tohave some set of standards so we can inte-grate the efforts,”said Goddard. ■
Internet gateway planned for neuroinformatics data
▲Haim Watzman,JerusalemA recalculation of Israel’s spending onresearch and development (R&D) showsthat it spends 3.5 per cent of its grossdomestic product (GDP) on civilianresearch — more than any other OECDcountry. The figures are the first in Israel toinclude accurate data on spending by smallsoftware companies in Israel.
The new picture comes from a surveycarried out by Israel’s Central Bureau ofStatistics. Previous assessments had placedthis figure at just under 3 per cent of GDP(military R&D figures remain classified).Such estimates for high-tech R&D spendingwere based primarily on information fromlarge companies, explains Simcha Bar-Eliezer, the statistician responsible for thesurvey.
Shlomo Herskovic, the director ofplanning and information for the Planningand Budgeting Committee of the Council
for Higher Education, which oversees andallocates the country’s higher-educationbudget, says the new figures change theoverall picture of Israel’s spending onresearch.
The survey shows that 21 per cent of thecountry’s civilian research spending is at theuniversities. Herskovic says the data couldbe used to argue that the universities shouldget more research funds. Given the extraspending in software companies, thegovernment is now responsible for only 8 per cent of civilian research expenditure.
The high level of research spending inthe software sector could also haveimplications for computer scienceprogrammes at universities. Even currentplans to increase the number of computerscience graduates by a factor of five over thenext three years may be inadequate tosupply software firms with the professionalsthey need, says Herskovic. ■
Software spend boosts Israeli R&D
Head first: brain data, such as this MRI scan, willsoon be brought together on the Internet.
SPL
© 2000 Macmillan Magazines Ltd