Provided by Christ Notes
Bible Search

If you can't measure your goals, then don't

Posted on March 14, 2013

"We realised that the price per analysis is the only standard which can be used to compare the laboratories"
The above is a quote found in the last issue of "Kemivärlden Biotech" and is attributed to Leif Nordqvist, project leader at the Swedish Competition Authority.
It strikes me that it is more and more common that the drive to have measurable goals gets more focus than if the goals are relevant. There was an excellent article series in "Dagens nyheter" a while ago about the state of the Swedish health care system which touches on similar things, easily quantifiable goals are allowed to control the decisions instead of the real quality.
Sure, quantifiable goals are good, but if you cannot measure the real goals, then don't measure something else and pretend that is your goal!

Z-score in Excel

Posted on August 30, 2012

I recently made an Excel udf to calculate the z-score (z-factor or what you would like to call it...) of a HTS plate. If you dare, you can download it here or if you prefer, copy the code below and paste into the VBA editor. The function is tested on Excel 2011 on Mac, but it should work at least with Excel 2010 and I believe also Excel 2007.

Function zscore(pos As Range, neg As Range) As Single
posmean = Application.WorksheetFunction.Average(pos)
negmean = Application.WorksheetFunction.Average(neg)
posstd = Application.WorksheetFunction.StDev(pos)
negstd = Application.WorksheetFunction.StDev(neg)
zscore = 1 - 3 * (posstd + negstd) / Math.Abs(posmean - negmean)
End Function

Biological activities database schema

Posted on October 14, 2011

In my day job I'm currently redesigning our database for biological activities of chemicals in preparation for our next big project that will hopefully start with a HTS within this year.
From the experiences I/we have had of our last system, I would like this one to be able to capture the raw data better, and then we would have software to calculate and visualize the end points (like percentage of control, fold change, EC50, etc.).
My idea is to base this on plates (with a generous definition of a plate that could also hold for example a set of cages with animals...) and experiments. I have attached an example DB schema to this post, click the image to enlarge to a zoomable version, or click here to download it as an MySQL Workbench document. Each well in the plate is marked as being positive or negative control, empty, or holding a test compound. There should also be information about everything that is put into it, such as cells, cofactors, stimulants, blockers, etc, and finally, there should be readouts, with the possibility to record time series.

If you have any comments, like the setup, or have ideas as to how this approach could be improved, please let me know in the comments.

Chemical structure editor

Posted on October 7, 2009

Thanks to a contest held by Metamolecular I started to think about what I believe are the most important functions for a chemical structure editor. (I use their product ChemWriter in a project of mine that hopefully will be available for public beta testing in a few months.)
Anyway, my main cirteria is ease of use! I like the editor to have simple keyboard shortcuts that are intuitive so that I remember them.
The first runners up is the looks, the editor should produce nice looking structures, preferably with the option to make custom labels that keep their chemical interpretation.

SpotSmart - Substructure search on your Mac

Posted on September 25, 2006

I have made a perl-script to enhance the excellent ChemSpotlight with substructure searching. The script in itself translates the output from mdfind into a valid smiles-file that could be searched using obgrep.

Update: A new version of the script solves an issue with long SMILES

More substructure searches

Posted on September 19, 2006

I have been using a customised version of the Lab Notebook from Macs In Chemistry as my labbook for several years now. I have however never really liked the substructure search in it. I don't want to first create a database and then search. When I'm looking for somehting I'm looking for it now... Besides, the fact that it uses ChemFinder that's still only available for Classic makes it even less attractive.

I have lately discovered the beauty of Open Babel and I figured that it's substructure search program obgrep should be possible to use to search in the Lab Notebook.