Lecture Notes: The Tissue Microarray Data Exchange Specification.
In: Specimen Processing Workshop. Sponsored by the
Epidemiology and Carcinogenesis faculty, NCI.
February 24, 2003, Rockville, MD.

The Association for Pathology Informatics has recently produced
version 1 of a community-based Tissue Microarray Data
Exchange Specification.

Like all other contemporary data exchange specifications, it
is built in XML.

It is extremely easy to implement.

Anyone can use it (it's Open Source).

I am not going to describe the specification because two
years of experience has taught me that the acceptance of the
specification is completely dependent on one issue.

WHY DO WE NEED A TMA DATA EXCHANGE STANDARD?

Basically, I just want to answer that question.

DIRTY LITTLE SECRET ABOUT TISSUE MICROARRAYS (TMAs)

People who do TMA research don't have any practical way to share
their data.

There is, for all practical purposes, no such thing as a tissue
microarray file.

When someone publishes a TMA paper, the TMA data set that supports
the paper's findings are not made available to the scientific
community.

When TMA slides from a TMA block are distributed to different
laboratories, there is no general strategy for re-assembling
the data from the different labs into a merged TMA file.

There is no community-accepted format for TMA data that allows
TMA data to be linked to other biological and medical data.

THE SCIENTIFIC COMMUNITY NEEDS A TMA DATA EXCHANGE STANDARD.

If we don't have a standard, we won't get much value out of this
new TMA technology.

Instead, we'll just get a lot more of what we have now:

TMA papers that consist of assertions that nobody can test
or build upon.

WHAT'S WRONG WITH THAT? Hasn't it worked forever and isn't it the
only practical way of doing science?

1. Editors won't put up with it much longer

Community Standards for Publication-Related Data and Materials.
Feb. 25, 2002, National Academy of Sciences. Washington, D.C.

2. NIH won't put up with it much longer (see NIH draft statement
for data sharing)

3. When you deal with terabytes of data, the data set becomes
the publication and the vehicle for scientific progress.
Journal articles are just little satellite editorials that describe
the data set.

EXAMPLE OF A DATA EXCHANGE STANDARD - THE LETTER

Sept 1, 1998

Jules J. Berman, Ph.D., M.D.
Program Director, Pathology Informatics
Cancer Diagnosis Program, DCTD, NCI, NIH
EPN - Room 6028
6130 Executive Blvd.
Rockville, MD  20892

Dear Dr. Kildare,

Blah blah blah....

Sincerely,

Jules Berman

WHY DO WE HAVE A STANDARD DATA EXCHANGE STANDARD FOR LETTERS?

Because a letter, by definition, is written for someone OTHER THAN
the letter writer.

If the letter were written for the letter writer, we wouldn't need
to learn how to write a letter.

If you're exchanging data and you expect anyone who receives the data
to understand the data, you need to have a data exchange standard.

If you're doing TMA research for yourself, you don't need a TMA
data exchange standard.

DIFFERENCE BETWEEN A DATA SPECIFICATION AND A DATA EXCHANGE SPECIFICATION

The data specification tells you the format (data architecture)
for that you must use to store and access your data.

The data exchange specification is simply the format you map your data
into before you exchange it with somone else.

Example:

Wordperfect files

Word files

RTF (rich text format) files

DIFFERENCE BETWEEN A DATA EXCHANGE STANDARD AND A DATA EXCHANGE
SPECIFICATION

Specification is a thorough description of how to do something
using standard descriptive techniques (in this case, XML and
ISO-11179 meta data)

A Standard has been approved by an organization that creates
and maintains new standards

THERE ARE A LIMITLESS NUMBER OF EXPERIMENTAL AND CLINICAL DATA
ENTITIES THAT NEED TO BE DIGITALLY EXCHANGED

protein arrays, gene expression arrays, EKGs, pathology reports,
tissue bank records, etc.

There just is not enough time or committees to do the job through
the Standards process.

It makes much more sense to get community support for data
exchange specification that conform to general standards for
exchanging information.