Our Next Scientific Database in Collaboration with Caltech: The Distributed Database of Electron Micrographs (DDEM)
Dr. Davi Ortega (PHD in physics), author of the Electron Tomogram Database at Caltech, has taken the time to provide an update on his work anchoring scientific records into the FLO Blockchain using the Open Index Protocol. In his update Davi tells us:
Since January The Jensen Lab at Caltech and Alexandria have been working on the next scientific database anchored in the FLO blockchain using OIP: the Distributed Database of Electron Micrographs (DDEM).
In 2018 we learned a lot during the design, building, and implementation of the ETDB. In particular, we understood that each ETDB “research.tomogram” OIP record would be much richer if certain fields of the record were filled with identifiers of other OIP records.
This seems easy to explain by example. In the OIP research.tomogram record “d9f5fc” (https://etdb.caltech.edu/tomogram/d9f5fc), the microscope used to take the tomogram is listed as “Caltech Polara”. This is not a very informative field, in particular to people that are not familiar with microscopy. Ideally, we would like to have a lot of information about this microscope. We would like for people to know that this tomogram was taken with a FEI G2 300 keV field emission gun microscope with a Gatan imaging fliers and K2 summit counting electron detector camera — this is the Caltech Polara. Obviously, we should not include all this info, repeatedly, over and over, in each of the OIP “research.tomogram” record.
Instead, we would like to have an OIP “research.microscope” record that lists the Caltech Polara specifications, far more detailed as the description above. When publishing the OIP record, we would receive the unique FLO txid identifier and then use that identifier to fill the field “microscope” in the “research.tomograms” records. This is true for many other fields: Microscopist, Aquisition software, processing software and etc… We recognize that this is something universal to any database using OIP and living in the FLO blockchain.
When we started to build the DDEM, we noticed that there are many types of OIP records that we would like to have: sample, protocol, equipment, scientist, laboratory, university and etc. However, OIP records are defined and hardcoded in the oipd: the official OIP daemon written in GoLang and maintained by the OIP working group — OIPwg.
In the current version of OIP, each project would need to fork the repository and implement the record types in the code. This would generate an ecosystem with multiple versions of OIP mostly incompatible.
The Power of Decentralization
This is what we have been up to in Q1. In the past few months, we have revamped the OIP logic to implement dynamic record types. Instead of having the record type definition hardcoded in the OIPd, we only have one record type hardcoded, the type “template”.
In the new OIP version, with a predicted release date for Q3 of 2019, projects will use the “template” to publish on the FLO blockchain the definition of the record type they want to use. Anyone can use any published record type.
As usual, this solution creates more complicated, fun problems to approach.
Relationships Between OIP Records
One of the fun problems we have been actively discussing is how to build a web of relationships between record types. For example, let’s suppose that Alice just graduated with a bachelor in science with a major in mathemagics from ACME University.
Alice has an OIP FLO profile and so does the ACME University.
We are working on the logic behind how ACME University would let the world know that Alice has this degree.
Right… the problem is that the logic behind the relationships between OIP records must be flexible enough to let ACME University tell the world Alice has a degree in Mathemagics, and to let Alice tell the world that she is in a band “Alice and the FLOers”, that the “Alice and the FLOers” just signed a contract with “Bob’s bar” to play a concert next Thursday. It has to be flexible enough to let the world know that the Jensen Lab is a laboratory at Caltech, that has sponsorship from NIH and HHMI, is in collaboration with Alexandria to build the next version of OIP and a team of postdocs and graduate students taking electron tomograms and micrographs of bacteria and environmental samples, and that each of these postdocs and graduate students received their degree from XYZ universities and … phew…
As you can see, we have big plans for OIP and FLO… and we are designing and building the tech and logic needed to make FLO the index cabinet of the world’s library.
This summer I will be heading to The Netherlands to expand the ETDB with a second university, Leiden University, the oldest university in the country and one of the most prestigious in Europe.