A donut with no jam – What use is a database with no data?
How Scientists don’t like boring jobs but if they can’t visualise data they can’t make the right decisions
In my recent travels to numerous Pharma and Biotech companies engaged in early phase drug discovery, one of the recurring themes that has struck me is the sheer volume of data that is being generated for them in the main by CROs. This data then has to be squeezed into the company research database which may be one of a very limited number of commercial ones, maybe most notably Dotmatics, or proprietary. Unfortunately, the data often comes from a range of different CROs in multiple formats and re-formatting the data for the database can be very time-consuming and hence nobody wants to do the job - but a database with no data in its pointless whereas once the data is in you can do so many exciting things like exports to support IVIVE or Human Dose predictions.
Here are just a few quotes that I have written down over the past couple of months:
The biggest issue is just dealing with data
Honestly for us, I would say it [dealing with data] is the biggest challenge to discovery when everything comes externally and that must be the case for a lot of companies
I spend a lot of my time re-formatting data for Dotmatics or Chemists.
The process of turning raw data from a mass spec to something viewable is massively valuable. If people can’t visualise data properly, how can they make the right decisions?
It appears to me that Scientists want to spend a lot of time generating data but hate the database side of things but without it how do you compare study data over 12 months old for instance?
Visualise Your Data
Manual data handling of course is error-prone and the more there is, the worse it gets. A recent example I have heard about was at a Biotech where they just found out that the wrong parameter had been uploaded for the last 6 months from an ADME assay. This was spotted when all the historical data was extracted and visualised demonstrating two important issues
A need to automate for accuracy
A need to visualise to see outliers or errors
Another example from a different Biotech revolved around when they extracted CYP data and realised no one was using it as the CYP risk never shifted on any project. In fact some projects never had a risk and screened extensively for no reason! This nicely illustrates a few further points
There is no point generating data for the sake of it
Analysis of trends is important
Data should enable decisions
I’ve highlighted on here before the advantages of integrating scientists from Compound Management, Chemistry, ADME and DMPK and one key way this can be done is through a common software platform. I have been involved in the design and implementation of solutions for automation of the whole cycle process from sample to upload for microsomes and other assays incorporating all the key steps such as sample lists, data processing, uploading and extracting.
The whole process from sample to database needs to work in a joined-up way and by making things communicate, people are empowered to focus on the things that really count and push projects forward.
If any of the issues discussed in these articles are of interest to you, please feel free to contact me directly on firstname.lastname@example.org for further information.
Especially get in touch if you are interested in tailoring CRO data packages to meet your database needs through standard templates and software support.