Database Design Problems

[SamLis]

Hi,
I watched the youtube movies, very interesting.
I tried to make the "business facts", well, I made a description of the attributes and how they are linked/dependent.

thanks for the help,
Sam

 

[jdraw]

A couple of articles:
1) local article on current cancer research

Ottawa doctor gets cancer research funding

Norman Jack with Alison Sandor
Thursday, October 11, 2012

The Terry Fox Foundation is awarding nearly $7.5 million to an Ottawa doctor for his work on cutting-edge cancer research.

Dr. John Bell and his team at the Ottawa Hospital have developed a special virus that, when injected into a tumour, can attack and kill the cancerous growth without the need for invasive surgery.

Dr. Bell says the funding will be used to try to understand how the virus works so they can increase the success rate of the treatment.

2)
This is an abstract of an experiment with nanobodies. (from http://www.ncbi.nlm.nih.gov/pubmed/22461363)
If you read this can you comment on the "general approach" and how it relates to your research. I'm interested in the "Things/entities" involved and also how well your data model could handle the approach in the abstract. If your situation is different, could we modify the general text to use your terms and get a paragraph that describes a typical experiment in your research (just to get the pieces well defined).

Abstract
RATIONALE:

A noninvasive tool allowing the detection of vulnerable atherosclerotic plaques is highly needed. By combining nanomolar affinities and fast blood clearance, nanobodies represent potential radiotracers for cardiovascular molecular imaging. Vascular cell adhesion molecule-1 (VCAM1) constitutes a relevant target for molecular imaging of atherosclerotic lesions.
OBJECTIVE:

We aimed to generate, radiolabel, and evaluate anti-VCAM1 nanobodies for noninvasive detection of atherosclerotic lesions.
METHODS AND RESULTS:

Ten anti-VCAM1 nanobodies were generated, radiolabeled with technetium-99m, and screened in vitro on mouse and human recombinant VCAM1 proteins and endothelial cells and in vivo in apolipoprotein E-deficient (ApoE(-/-)) mice. A nontargeting control nanobody was used in all experiments to demonstrate specificity. All nanobodies displayed nanomolar affinities for murine VCAM1. Flow cytometry analyses using human human umbilical vein endothelial cells indicated murine and human VCAM1 cross-reactivity for 6 of 10 nanobodies. The lead compound cAbVCAM1-5 was cross-reactive for human VCAM1 and exhibited high lesion-to-control (4.95±0.85), lesion-to-heart (8.30±1.11), and lesion-to-blood ratios (4.32±0.48) (P<0.05 versus control C57Bl/6J mice). Aortic arch atherosclerotic lesions of ApoE(-/-) mice were successfully identified by single-photon emission computed tomography imaging. (99m)Tc-cAbVCAM1-5 binding specificity was demonstrated by in vivo competition experiments. Autoradiography and immunohistochemistry further confirmed cAbVCAM1-5 uptake in VCAM1-positive lesions.
CONCLUSIONS:

The (99m)Tc-labeled, anti-VCAM1 nanobody cAbVCAM1-5 allowed noninvasive detection of VCAM1 expression and displayed mouse and human cross-reactivity. Therefore, this study demonstrates the potential of nanobodies as a new class of radiotracers for cardiovascular applications. The nanobody technology might evolve into an important research tool for targeted imaging of atherosclerotic lesions and has the potential for fast clinical translation.

I have added some comments to your word doc and saved it as new doc in the zip.

 

[SamLis]

Hi Jdraw,

I adapted the word file again.

you quoted 2 papers, you must understand that this papers, are made up by many experiments, to bring this kind of organization in the database is not my goal.

I you recall the filemaker form, it says @ the bottom, Applications : Western Blot, immunoflorescence with a checkbox.
I want to expand that a bit by: ok the checkbox Western Blot says yes,
who did that experiment? when? Results? By results I mean did it work, and a link to lab notebook.
People can then check in the notebook the real details.

-Sam

 

[jdraw]

Yes I understand the source of the materials I posted. The first was from the local news today. I just posted it to say that the kind of research you are talking about is in the news.

The second article was more of the jargon you deal with. I wanted you to take a paragraph or two of what most of us (including the guy at McDonald's) would consider scientific gibberish and collapse it into something the average person might understand.

The article dealt with terms that we could hopefully relate to your experiments, tags, techniques, epitopes, species reactivity etc. We're trying to build a picture to help us understand these Things/Entities and how they relate. The actual attributes and values you use to flesh these out will be in your terminology.

I don't know if you created form(s) similar to the filemaker form for each of your needs, so I copied it and posted a copy. Anything on a form will reside in a table somewhere, hopefully you can map some forms to your entities/tables.

I have to go out for while.

 

[SamLis]

Hi,
I put textboxes on the form, these are the attributes that both db's share, they have some I don't need and other are lacking.

Like the TagPosition and Tag :

But for each Nb : either C or N terminal + one tag
as mentioned before :
Nanobody 1 C-Term 6His
Nanobody 1 C-Term GST

is the same nb with a different tag, so different records

 

[jdraw]

Do you have some sample data? Can be fictitious data but must adhere to your general business facts for good data. It's also good to have some data that violates some criteria to ensure your edit/validation routine catch them.

You need some good data for all your tables, and as you go through your business facts you're likely to find some anomalies. This will force you to determine if the data is bad, or if the model is bad, or if the business fact needs clarification.

Are you aware of http://www.expasy.ch/vg/index

 

[jdraw]

There is an article at the following site that indicates a free (for academic purposes) ms access database and a .net executable. This may be useful to you. You would probably email the person in the link and request info and/or a copy or advice on how you might proceed.
http://intl-jla.sagepub.com/content/16/1/82.full

There is an Access database template at this site. Not specific to nanobody, but with some related info and a writeup about the database.
http://www.ibridgenetwork.org/uctech/salgia-thoracic-oncology-access-template

There is also a detailed database (in filemaker) at
http://www.boxit-labstorage.de/BOXIT_intro.html
It seems quite elaborate and widely used. You may want to pursue such a database and filemaker rather than build something in Access. Building a comprehensive database is not a trivial task. Buying an established product may be the simplest and easiest path to a useful, sharable productive database application.

Good luck with your project.

 

[SamLis]

Hi,

I know there are good Software Apps out there, however they are quite expensive, since we are a big lab ~ 40 -50 people, it would be very costly to buy licences for everybody.
Anyway I added some data, and indeed some problems did arise ...
see the doc.

 

[SamLis]

Hi ,
I checked this Box-It app, but it didn't stop me or gave an error when I entered 2 times the same item? I thought this was the basic idea of a good relational database?

Greetz,
Sam

 

[SamLis]

Hi,
I adapted my tables and relationships to solve the issues I had before, well I think I did?

Sam

 

[jdraw]

Hi,

I know there are good Software Apps out there, however they are quite expensive, since we are a big lab ~ 40 -50 people, it would be very costly to buy licences for everybody.
Anyway I added some data, and indeed some problems did arise ...
see the doc.

I noticed that your TblAntigen data and your tblEpitopeGeneral have the exact same data. As I tried to point out earlier, it is critical to define and understand exactly what each entity/table represents. Table and relationships are key to database.

If you are a large lab, and are research oriented, then I would suggest a commercial software application may be your best choice. When you buy a commercial product there is technical support, typically a user group/community, and current software fixes/mods to problems. The cost of such a package may not be so expensive when you consider the efforts to design an application, purchase software, develop, test and maintain the application, create a database support group, establish standard practices and disciplines to manage all aspects of data, design, testing, modifications etc.

What do you do currently as far as experiments, naming conventions, potential patents, detail data collection and maintenance?

Building something for personal use is one thing, but putting a research organization's business into a database application is another level of risk that requires more discipline and support. You will find the software acquisition costs may be very small in comparison to the effort and resources for an application to support a 40-50 person lab.

 

[jdraw]

Hi ,
I checked this Box-It app, but it didn't stop me or gave an error when I entered 2 times the same item? I thought this was the basic idea of a good relational database?

Greetz,
Sam

I did not check into the actual database. Just that it was available.

I would not base any judgement on a few samples in a test. There seemed to be a lot of orgs using the application as i recall.

I would not discard this application at this time.
You may want to do some research on
identifying needs/requirements for selecting a software application

 

[jdraw]

I have to go out for a few hours and I have not looked at the last database you attached. .

 

[jdraw]

I have just looked at your latest database.
I have revised the relationships layout so no lines cross.
I have removed(spaces) field names with embedded spaces.

tblNanobodies should Not contain
ExpressionVector, it should be ExpressionId
Tag, it should be TagId

I'm not yet convinced that Aliquot... should be in tblNanobodies.

tblNanobodyFix (whatever that is) should not contain
*****NanobodyName, it should be NanobodyId.
Epitope, it should be EpitopeGeneralId,
EpitopeDetail, it should be EpitopeDetailId
SpeciesReactivity, it should be SpeciesReactivityId
Production, it should be ProductionId

***Decide where nanobodyName belongs either tblNanobodies or tblNanobodyFix
Pick one or the other only and in the related table use NanobodyId


tblProduction should NOT contain
antigen, it should be antigenId

Remove all lookups from your Table designs
--techniqueId in tblExperiment
--userId in tblExperiment
--ExpressionVector
.......there are more

see
http://access.mvps.org/access/lookupfields.htm

I have attached a revised version of your database.(we should use a zip)
Can we work with this until we get the structure set up? You can always move it to 2010, but I can't use it.

We need some test data.
Can you concoct some sample data that we could put into the current "model" and see if there are adjustments required?

 

[SamLis]

Hi,
I updated the db, removed some fields.
I started from a physical situation, translated into the db.
Since I removed the lookups, I used ID's to fill in for now, that's why there is this scheme attached.
I focused now on the nb part , not on the experiments.

thanks for the help so far,
feel like we are getting there

 

[jdraw]

What is difference in NanoBody vs NanoBodyFix?

In a normalized structure, the PK uniquely identifies a record. There are no partial dependencies.

My concern was the appearance of NanoBodyName in 2 distinct tables -- that violates Normalization rules. If you saw something in an article I sent, it may be a mis interpretation???
I think we should sort this one out before moving too far.

I'm not sure where you are. I am in EDT Time zone. Today I am going to my cottage to close up for the year, so won't be online much.

 

[SamLis]

Hi,

I'm in Belgium so CEST timezone.
While I was filling in the database I realized that if you know the name, you know the epitopeDetail, SpeciesReactivity, Production.

So I stored that Data in a separate tblNanobodyFix.
So I avoid that people can select a Name and an Epitope, no if you select name you already limit yourself to an epitope? Does it make sense?

In Nanobodyfix I stored things that cannot change, one name -} Epitope, ... .
In tblNanobodies you have additional info like Tags, that can change from one entry to the next.

in my example tblNanobodies you have NanobodyID 1 & 4
It is the same nb in relation to Production, Epitope, ...
but we changed the tag, ExpressionVector.

In the Word file I described the situation.

THanks,
Sam

-Sam

 

[SamLis]

hi,

maybe the confusion comes from the fact I forgot to change the tblNanobodies NanobodyNameID into NanobodyFixID?

Simply put (as you can see in my example ) :

The combination NanobodyName, EpitopeDetail, SpeciesReactivity, ProductionID is not unique to a nb.
it's this combo (tblNanobody fix) + TagPosition, Tag, Expression that make a nb unique.

-Sam

 

[jdraw]

You are 6 hours ahead of me, so it's about 11:06 PM at your end.
I'll look at your tables and make a couple of adjustments as you suggested.
When I look at tblNanobodies I notice that
TagId has the same value as ExpressionId <----Is that just coincidence in your test data?

Alao, I will removew the embedded space in ImmunizationDate in tblProduction

I'll look more at this later tonight (my time) and comment/question if necessary.

 

[jdraw]

I did look at the database tonight. I created a query to show some of the info you could retrieve based on the top part of your powerpoint sample.

I have updated the db with the query in the attached zip.

There is still confusion, on my part, with NanoBody and NanoBodyFix.