bhiWay

i’m not going to post about diy bio here anymore, the  http://www.diybiosea.org/ blog will handle that.

oh and i just did a wordpress update, worked great

Last night the Seattle DIY biologists met for a second time, this time there were seventeen of us including myself, a first timer.  The group trickled into Dan Heidel’s house in Phinney at around 8pm.  Dan began by telling us about his project to set up lab space in a commercial space he rented in south Seattle which he is making open for use for serious projects that any of us would like to undertake.  He has been buying up equipment and it is really starting to come together, but he feels that if someone did start a project now chances are at least one piece of necessary equipment would be missing.  Not to mention consumables which he has none of.  Needless to say his effort, called Seattle Open Bio Labs, LLC, is an amazing step forward in organizing and a very generous gesture to share it with the group.  Dan has been buying equipment to create this environment where he hopes to create a community of Open Science projects.  He hopes to encourage all to keep their lab notebooks online and open as transparency and open collaboration is a powerful driving force behind innovation.  Of course if someone has a commercial goal in mind Dan is willing to listen about specific cases.  The question of where to find funding for running this operation and supporting projects is a unanswered question, and for the time being individuals would have to fund themselves.  The discussion of Dan’s lab answers the main question from Sandra Porter’s post (see 1st meeting) about where to practice DIYBio.  (Sandra was unable to make this meeting).  Still there remain some todo’s before starting, such as contacting the fire department for advice on any permits and addressing concerns of the wary (Dan does have Liability Insurance for the lab space).

After a significant amount of time we moved on. Dan does talk a lot, but not in a bad way.  We discussed project ideas from various other members of the group.

1. Lifesuit to help paralyzed people walk (see http://theyshallwalk.org/ for more)
2. Growing neurons to generate solenoids, using growth factors and maybe brewing GFP beer!
3. Produce our own enzymes to share the stock with the international DIY Bio community
4. in vivo DNA synthesis, using light as input to dictate to the cell the sequence of DNA.  This would allow you to change the code of the cell while it was running.
5. Setting up DIY Computational Biology (smallest amount of hurdles on this one) BeBoBio
6. Making fuels from microorganisms
7. Geothermal cycling for speeding up growth
8. Nanosilver
9. Self stable ecological systems.
10. Bacteria that can metabolize polyethylene; Dan said Pseudomonas Originalis which smell like grapes and plague burn victims may be capable
11. Bio Weather Maps

Atendees:

Randy Hall, Kris Ganjam, Monty Reed, Dan Heidel, James Yang, Alec Nielsen, Tyler Casey, Matt Crowley, Michal Galdzicki, Scott Mason, Tracy Tucker, Max Berry, Ron Shevuah (4 names missing)

[I am missing some people and project ideas, please contact me if you were there but i didnt jot down your name, sorry.  Some peoples names i got from the RSVPs so please if you weren't there and i listed you above I appologize. ask me to correct it]

Liu:

Using ontologies such as MGED ontology (MO), NCI Thesaurus to annotate microarray data from GEO.  They mapped MO and NCIt.  The BCM-CO prototype contained 1200 classes and 5500 synonym terms, these were used to find GEO descriptions of breast cancer single channel arrays.  They discussed the NCIt with regards to whether or not the terms found in the descriptions were found or not in the thesaurus.  Especially the compositional terms were hard to find. The indexed data were used to rerieve potential analysis sets.

Tu:
The NDAR repository for data from NIH funded studies of autism.  The features such as age, verbal and non-verbal IQ, ADOS and ADI-R scores.  They surveyed literature of autism and extracted terms and relationships to build ontology.  They built an ontology of autism phenotypes and define the phenotype in PATO terms and BIRNlex information is extended.  SWRL rules were used to define autims phenotypes in terms of data for data analysis.  This coding was written to conclude a subjects phenotype by assessing whether the data code contained the code.  Then they query the data set with automated inference of phenotype abstractions.  They conclude that you can do data inference, not just annotation. They state that the information model is part of the ontology, and claim generalizability to broader clinical data.

Detwiler: Regular Paths in SparQL: Querying the NCI Thesaurus Native OWL representations are obscured.  The NCIt browser simplifies the view of the ontology graph.  The underlying OWL representation is far more complicated: OWL class definitions which link properties to the class have an intermediate “restriction”.  The Gleen extensions allow to define regular expressions, as an extension implemented as a plugin.

Sharp (presented by Olivier Bodenreider): A Framework for Characterizing Drug Information Sources
Drug information sources are varied and none are comprehensive. The drugs have information about the such as Pharmacy, Chemistry, Biology, Clinical medicine (here excluded pricing/ packaging).  These 4 domains represent the groupings by which information is organized for evaluation (to check whether that resource covers that area), this data was PCA’d and the domains cluster within the 1st two PCs.  DailyMed, WHO-ATC, UMLS, DrugBank found to be the best.

Cook: Bridging Biological Ontologies and Biosimulation: The Ontology of Physics for Biology

Biosimulation semnatics.  1st biosimulations models are hand crafted, the code is formal, but the meaning of those models is not.  The SemSim ontology maps the simulation code to refrence ontolgies.  A semantic map between the computational model and the physical model.  The physical properties and physical dependecies are the key to the OPB structure defined by the four types of properties: flow, displacement, force, and momentum.  The dependencies are the relationships between these kinds of properties.  The OPB can serve to map the physical relationships within the mathematical representations of biological processes.

Piccolo: Somatic Mutation Signatures of Cancer (3rd prize for student paper)
Classification of cancer motivates this work.  Cancer types as in location and histology.  Aim to differentiate between cancer types. Applied the Vogelstein model as a guide. Catalog of Somatic Mutations in Cancer (COSMIC), used as source of studies which typically analyze single or few genes at a time.  They picked the mutations which have a proportionately high contribution to cancer.  They performed a type of machine learning (?) and then clustering on the represenatative vectors, using manhatan distance and hierachical clustering.  The somatic mutation differences between colorectal adenoma and carcinoma are similar, according to their somatic mutation molecular profile.  However, breast cancer is distant.

Musen

Reported on the use of BioPortal 2.0 as a community based access point to ontologies and actually link to the data they are ment to describe.  The community tools include, browsing; including visualization (from P. Storey, Victoria, B.C.), and comments, linking (resources), voting (to reach consensus), the inclusion criteria are loose (related to biomedical research).  Also provides access to mappings between ontologies.  The projects and notes pages provides a place for testimonials of use and challenges, questions, and critisisim.  The ontology metadata provides descriptions of the ontology itself.  API and documentation are also available.  Mapping are manual at the moment, but Musen suggested that prompt mapping uploads should be possible.

Shankar (and team) -

Talk + Demo of TrialWiz. from ITN at Stanford.  Multiple applications integrated using an knowledge management (OWL, OWL/SWRL) and data management (DB) framework.  SWRL used for mapping.  Epoch ontologies.

BIRN no show?

Dan Masys covered both Clinical and Bioinformatics

Clinical Informatics

CDSS for Providers

• Warfarin, asprin studies showed improved adherence to guidelines

CDSS for Patients

• Benzodiazepine: tailored intervention gets patients to stop meds (easier to stop then to start patients on meds, hahah)
• Hazardous drinking in college: online information intervention had impact to lower drinking, was sustained >12mo
• Automated test result reporting: knowledge is power (increased overall satisfaction)
• Quality of life and Needs for care: structured patient-provider dialogue around patients world view, helped relate patients values..

CDSS – No diff reported

• Matheny – med lab monitor
• Hicks – BP management
• Tamblyn
• Thomas – registry – audit
• Harari – health risk appraisal
• Hansagi – ER info -> primary care provider

PHRs

• Grant – PHR links EMR for type-2 diabetes; previsit use of PHRs had impact on med change (highly motivated cohort)

Telemedicine

• 10 new RCTs; 3 for hypertension
• Green, BB; Santamore, WP; Madsen, LB – all on BP  (secure website w/+w/o pharmacy ) -> telemedicine equivalent to in person monitoring for chronic conditions; adds to already known findings
• Meyer, BC in Lancet Neur. – 2-way audio viedeo v. telephone for thrombolytics; more accurate stroke decision making; useful way to do things for specialists.
• Shea S – diabetes  care by telemedicine for low income populations; BP, LDL, HBa1c all got better; telemedicine effective but no cost analysis (comments included cost benefit to family not just provider)
• Ellison – post-operative robotic rounds; Morbitity and length of stay was equivalent; patient satisfaction was also equivalent; implies: doctor bedside = robot bedside

Telemedicine No diff

• Dansky – heart failure
• Leimig R – transplants

Practice of Informatics

• Love – mining EMRs for CTSAs
• Bareznicki – data mine community pharmacy med records to improve asthma managment
• Meystre SM – automated problem list w/ improved sensitivity; What is the utility of the problem list? This was an intervention to change clinical practice but pilot has challnages scaling up.

Bioinformatics/ Computational Biology

• Cooper (from UW) – Warfarin dosing; aim to add more common SNPs to the current two genes (VKORC1, CYP2C9) via GWAS.  What genotypes benefit from which therapy dosage?  Pathway interpretation of SNPs early SysBio approach to functional understanding.
• Castellanos-Rubio – Celiac disease; linkage with SERPINE2 2q33, PBY3/PPP6C 9q34; Complementary use of structural and functional information to find them.
• Chaussabel – Systemic Lupus: microarray approach which includes functional classification to boost relevance.  System scale appraoch combining quantitative data.

TOP TEN

10. Personal Genome Project

9. ONL Strategic Plan (Helth IT)

8. Mass and NV pass laws requiring encryption of personal data devices

7. 1st HITSP standards exchnage (NHIN)

6. AMIA Rockefeller Foundation Global eHealth Connection Conference

5. CMS Medicare Improvements Act – ePrescribing pays more.

4. Explosion of Molecular Data – 2nd, 3rd, 4th Personal Genomes; proteomics 1TB/ experiment; infrastructure strained; we’re behind the power curve

3. FDA Sentinel Intiative

2. NIH Open Access Policy

1. Obama

Mougin: Used an approach using mapping to find semantic errors in the NCI thesaurus.  One of the conclusions was that Pellet is slow; however that was disputed by the audience response.  The

Denny: Created a new terminology for clinical notes by building it from a training set of notes.  Parsed things which look like headers from the structured note.

Fung: RxTerm: Chopped up version of RxNorm: appears to be inteded for data entry of medications, as the original RxNorm is really exhaustive, as it has dosage and route information all pre-coordinated. RxTerm reduces the number of displayed choices as you type the 1st 4 or 5 chars.

Lunch meeting at Clark & Parsia. Kendall Clark and Michael Grove were very nice to have invited me to have lunch with them and to chat about the Semantic Web work that they are doing.  I got a tour of the office and discussed my project.  Kendall suggested looking at some Mike Smith’s modularity work.  Meeting Evren Sirin, Michael Smith, and Markus Stocker was impressive as well.