CINF ACS San Diego

1
Smart Laboratories to Smart Papers

• Jeremy Frey
• School of Chemistry, University of Southampton
  UK.
• Ubiquitous Computing and e-Research
• NeSC May 2005

2
Smart Places
3
Chemists and programming

• Many Chemists think that they can program
• So leave the systems to the Chemists

You still use FORTRAN!!
4
e-Workflow
Some Chemists can and leaving it to the computer
scientists can give you a perfect system for a
problem you didnt know you had!
What about that! His brain uses formal logic! No
wonder it took so long to get a result
5
Talk Workflow

• Introduction to e-Science the Combechem Project
• Pub / Sub
• Semantics for Smart but not Dark Labs
• Conclusions Publication_at_Source

6
The CombeChem Project

• Collect data with regard to how it could
  eventually be used
• Make sure the metadata is of high quality
• Record properly at source in Digital Form
• The Chemistry Lab
• People Machines working together
• End to End linking of data and information
• Publication_at_Source
• But then.... Who needs provenance? Bush, Blair
  The JIC, MI5, CIA Hutton 2004

7
Smart Laboratories

• Laboratory
• Environment
• Equipment
• Smart and legacy systems
• People (Smart People?)

8
He is charged with expressing contempt for
meta-data
9
Adapt SHG lab

• Set out to store the data coming from our
  non-linear laser experiment
• Use a database!

10
Digital Camera very wide spread high end
computing
11
Capture information from places you would not
want to put your eyes
12
Ti sapphire laser system - schematic
Seed laser - 1nJ, 20fs, 80MHz
Dazzler - arbitrary dispersion generator
second amplifier - pulse amplified to 3 mJ, 1 kHz
Pulse stretcher - 20fs - 400ps
Pulse compressor - 400ps - 32fs
Regenerative amplifier - pulse amplified to 0.5
mJ, 1 kHz
Output 32fs pulse _at_780nm, 3mJ _at_1kHz Focused
intensity 1015 W/cm2 E field 80GV/m
13
(No Transcript)
14
(No Transcript)
15
Monitor the semi-automated laboratory
Data Source
Data Source
Translator Service
Client
Client
Client
PDA
16
Use IBM Web Sphere Technology (MQTT) - an
asynchronous grid
Data Source
Data Source
Message Broker
Translator Service
Client
Client
Client
PDA
17
(No Transcript)
18
(No Transcript)
19
Yesterday in the lab
Temperature room, laser
Door interlock Motion Sensors
20
Student turned off a/c
Air Conditioning failed
21
(No Transcript)
22
What you have to do to get on to the BBC web site
Pub/Sub for Laboratory data using a broker and
ultimately delivered over GPRS
23
(No Transcript)
24
I love the convenience, but the roaming charges
are killing me
25
What are the people up to?

• We can capture the environment but need to
  capture the processes

26
We dont want to take this technology too
far. There are privacy issues
When we implant your pacemaker, we can, for a
modest additional fee, also implant your beeper.
27
Industrial support
Big block to publication_at_source if its not
digital, its difficult to share
critical data entry
28
Like cash machine electronic notebooks have taken
many forms we want one that adapts to context,
simple to use in the lab whilst doing actual
bench chemistry but linked to more resources at
the desk the pervasive lab book grid
29
Smart Tea Keep the people involved!
I just realized, Howard, that everything in this
apartment is more sophisticated than we are
30
Getting not just the what and how, but the why
Get chemists and computer scientists to
understand each other By Making Tea!
31
COSHHleverage off things we already have to do
32
Tablet version
33
Use RDF for both data and processes
34
A digital lab book replacement that chemists were
able to use, and liked.
35
(No Transcript)
36
Annotation_at_source
37
(No Transcript)
38
Scaling?
t 0.0003n28.487n R0.9998
39
Experiments on the Grid
Are we simply building a rod to beat ourselves
with? Is the user out there?
National CrystallographyGrid Service
No, the computers are up, Were down
40
Application







41
Sample Submission
42
Status Monitoring
CLIENT
NCS
43
Security and trust for experiments and data
44
Data Collection
45
Automatic Structure Solution

• Background process designed to adopt the Human
  Approach, using refinement indicators and
  structural knowledge
• Encorporates all Q peaks above a cut-off as C
  atoms
• Reject on basis of thermal parameters, adjust
  atom types accordingly iterate
• Hybridisation hydrogens from connectivity
  difference map peaks then fixed
• Usual crystallographic validation performed,
  -introducing chemical validation

46
Data Collection

• Metadata capture

47
Data Overload the Publication Problem
2,000,000
25,000,000
300,000
48
Separating Data from Interpretations
Underlying data
Intellect Interpretation
49
Simple Deposition
Metadata attached
50
An Archive Entry
ecrystals.chem.soton.ac.uk
51
Access to ALL underlying data
52
Remote control?
53
Several groups making and analysing the library
Administrative Domains transfer or share the data
National Archive
Research Group
Researcher
Research Group
Institution
International Database
54
(No Transcript)
55
SVG active graphics Links to raw data pulled
from the database
Follow the data processes recorded in RDF infer
the pathway/workflow
56
(No Transcript)
57
RDF

• Issues for real time information
• Inference and Context is very important
• Assertions may only be valid over a certain time
  period
• Information may be inaccurate and conflict
• Lots of potentially irrelevant data
• But still miss the important issues

58
(No Transcript)
59
First, they do an online search
60
(No Transcript)
61
(No Transcript)