|
You are here >
-Omics and Informatics
DRAFT
Genomics and Bioinformatics for Drug Discovery and Development
Mary Chitty mchitty@healthtech.com
617-630-1316
not being updated
Scope:
These sections focus on the use of genomics and proteomics (and
other -omics) in combination with bioinformatics, cheminformatics
for drug discovery and development, current and potential uses in pre-
clinical, and clinical trials, and the quickly growing, still
evolving role of molecular medicine. Biopharmaceutical
manufacturing is not a major focus, but the efforts involved in ramping up to
higher throughput and scaling in going from the R&D lab into clinical use
are not inconsequential.
Biotechnology
began as a means for producing biopharmaceuticals. It is
still in the process of becoming an integral part of drug discovery and
development. Use of genomics and proteomics is still primarily at the earlier
stages of drug discovery pipelines.
Key information resources
Biology
Genomics Information resources
Proteomics information resources
Bioinformatics information
resources
Chemistry
Combinatorial chemistry & chemical genomics information resources
Cheminformatics
information resources
Business information resources
Drug discovery and development Information resources
Pharmacogenomics
information resources
Toxicogenomics information resources
Clinical
trials
Molecular Medicine
Information resources
Redefining diseases
Redefining diagnostics
Bioethics information resources
Bio Terminology information
resources
Summary What do (and
don’t) we know now?
Genomics
We still don’t know what over half of
human genes do (functional genomics).
Simpler organisms, used experimentally, can help determine human functions (comparative
genomics).
Genomics is still at the earliest stages of the drug discovery and development
process.
Bioinformatics
Isn't just about sequencing anymore.
Pharmaceutical R&D is pushing the computational envelope trying to analyze
and meaningfully interpret very large quantities of extremely noisy biological
data.
Proteomics
But proteins do most of the work.
Humans may only have 30K genes. (Wheat and barley have more.)
Estimated number of human proteins:
With alternative splicing, post-translational modifications and differential
gene and protein expression, in different cells and tissues and during
different points in cell cycles AND varying in health and disease there
is still a lot to understand.
The most commonly expressed proteins are probably the least informative.
Better ways of measuring and studying rare proteins and transcripts are needed.
Chemistry
There isn’t enough matter in the universe to make all the possible
combinations of combinatorial chemistry
Chemical genomics uses chemical probes to aid both drug target identification
and
target validation.
Cheminformatics
Keeping track of and meaningfully interpreting all the variables gets even
harder with three dimensional protein structures, and four
dimensions plus (temporal- spatial aspects) and highly dimensional (protein-protein
interactions) data.
New algorithms are required.
More and better data is needed.
Drug discovery
We have an unprecedented glut of novel targets. Technologies
for validating targets haven’t caught up yet.
Drug development
Is increasingly
information- driven.
(How many drugs now have known mechanisms of action? 30% up from zero since
when?)
Linear handoffs between biology and chemistry need to become more parallel.
Compounds that will fail eventually need to fail sooner (reducing attrition)
to lower costs of later phase trials.
Target validation
(not just target suspects and target
identification) must reflect specified diseases, showing "reasonable
evidence" that target modulation reverses (or ameliorates) disease
phenotypes, not just suggest that a gene/protein "causes" a disease.
Weather prediction is a useful
metaphor.
Technologies
Metabolic profiling (high throughput assessment of tissue or cellular metabolites),
can create effective molecular phenotypes.
Integrating metabolic profiling with genomic and proteomic data for a systems
biology approach is promising.
RNAi (RNA interference) is being used for more
effective target screening and to reduce costs.
Using functional validation earlier in the pipeline process is
attractive.
Current functional validation strategies (such as antisense) will be used
at earlier stages if cost and throughput limitations are overcome.
Microarrays have promising clinical research applications, particularly in oncology. (So
do even more nascent protein arrays).
Automating, reproducibility, ramping up throughput and cost-effectiveness are
all serious challenges in moving technologies from the R&D lab into clinical
testing, and eventual FDA approval.
Pharmacogenomics/Personalized medicine
holds great potential – and peril, threatening further fragmentation of an
already fragmented industry.
But could also speed clinical trials, create more targeted, less toxic drugs,
and reduce costs incurred now by overtreating patients not needing, or unlikely
to benefit from specific drugs.
Toxicogenomics
Appears much closer to adding value than pharmacogenomics, reducing time
required for safety evaluations and perhaps lessening the likelihood of late
stage (expensive) tox surprises.
Molecular medicine
is already here (600+ clinical tests in GeneClinics, 300+ research), but
not for all patients or all diseases.
Business
It's nice to get paid on a regular basis.
Current increases in R&D costs and decreases in new NCEs and NMEs is not
sustainable in the long run.
The risk-adjusted cost per new drug developed under precedented targets came
to $700 million, according to Philip Ma, (of McKinsey, in August 2002 Nature
Drug Discovery Reviews). By comparison, the risk-adjusted cost per novel
target new drug came to $1 billion. Over time, as experience is gained with
targets in new protein family classes, the difficulty of developing newer
compounds will begin to diminish, and the value of working on novel targets will
become more favorable. Just as there are reasons to expect performance to
improve in target validation, improved returns on investments in genomics for
other applications should as well.
Genomics is now in an awkward adolescence. Disappointments have been great
because the tremendous potentials created unrealistic expectations. Even experts
failed to recognize, or chose not to focus on, the challenges involved in
implementing strategies with such profound implications. Maturation and
adjustments are an inevitable part of introducing new technologies into the drug
discovery and development process.
What Information
professionals can do
Recognize tough choices.
Focus, without putting all eggs in one basket.
Look for allies and potential partners (in house and outside) and ways to
leverage existing strengths.
Look for ways of integrating information and communicating between
"information silos".
Learn the language and terminology of different specialties.
The more you know, the more you can admit all you don’t know.
There is a lot nobody knows yet.
Help people articulate what would/might help – at a time when this gets harder
to express or figure out.
Working harder isn’t always possible. Working smarter – and with others –
probably is.
Expect "equity bets". Plan for ongoing, profoundly dynamic change.
Think globally. Act locally.
What it means
"A month (or two) in the lab can save you an hour in the [virtual or
physical] library." More true than ever.
Key Information Resources
Genomics and Bioinformatics (and beyond)
Journals and news
Nature, particularly
the front matter.
Nature
Biotechnology
Nature
Reviews Drug Discovery
Science,
particularly the front matter.
Business Week, Economist, Financial
Times, Forbes,
Wall Street Journal others? Which are the
most indispensable? Business
journals
Google News
http://news.google.com/
Search for specific topics.
Genome News Network
Databases Database
directories
Websites
NIH Roadmap http://nihroadmap.nih.gov/
What else? Please send me your nominations mchitty@healthtech.com
|
