Archive for the ‘research’ Category
Encryption of chemical information has not been a very common topic in cheminformatics. There was an ACS symposium in 2005 (summary) that had a number of presentations on the topic of “safe exchange” of chemical information – i.e., exchanging information on chemical structures without sharing the structures themselves. The common thread running through many presentations was to identify representations (a.k.a, descriptors) that can be used for useful computation (e.g., regression or classification models or similarity searches) but do not allow one to (easily) regenerate the structure. Examples include the use of PASS descriptors and various topological indices. Non-descriptor based approaches included, surrogate data (that is structures of related molecules with similar properties) and most recently, scaffold networks. Also, Masek et al, JCIM, 2008 described a procedure to assess the risk of revealing structure information given a set of descriptors.
As indicated by Tetko et al, descriptor based approaches are liable to dictionary based attacks. Theoretically if one fully enumerates all possible molecules and computes the descriptors it would be trivial to obtain the structure of an obfuscated molecule. While this is not currently practical, Masek et al have already shown that an evolutionary algorithm can reconstruct the exact (or closely related) structure from BCUT descriptors in a reasonable time frame and Wong & Burkowski, JCheminf, 2009 described a kernel approach to generating structures from a set of descriptors (though they were considering the inverse QSAR problem rather than chemical privacy). Uptil now I wasn’t aware of approaches that were truly one way – impossible to regenerate the structure from the descriptors, yet also perform useful computations.
Which brings me to an interesting paper by Shimuzu et al which describes a cryptographic approach to chemical structure search, based on homomorphic encryption. A homomorphic encryption scheme allows one to perform computations on the encrypted (usually based on PKI) input leading to an encrypted result, which when decrypted gives the same result as if one had performed the computation on the clear (i.e., unecnrypted) input. Now, a “computation” can involve a variety of operations – addition, multiplication etc. Till recently, most homomorphic schemes were restricted to one or a few operations (and so are termed partially homomorphic). It was only in 2009 that a practical proposal for a fully homomorphic (i.e., supporting arbitrary computations) cryptosystem was described. See this excellent blog post for more details on homomorphic cryptosystems.
The work by Shimuzu et al addresses the specific case of a user trying to identify molecules from a database that are similar to a query structure. They consider a simplified situation where the user is only interested in the count of molecules above a similarity threshold. Two constraints are:
- Ensure that the database does not know the actual query structure
- The user should not gain information about the database contents (except for number of similar molecules)
Their scheme is based on a additive homomorphic system (i.e., the only operation supported on the encrypted data is addition) and employs binary fingerprints and the Tversky similarity metric (which can be reduced to Tanimoto if required). I note that they used 166-bit MACCS keys. Since it’s small and each bit position is known it seems that some information could leak out of the encrypted fingerprint or be subject to a dictionary attack. I’d have expected that using a larger hashed fingerprint would have helped improve the security. (Though I suspect that the encryption of the query fingerprint alleviates this issue). Another interesting feature, designed to prevent information about the database leaking back to the user is the use of “dummies” – random, encrypted (by the users public key) integers that are mixed with the true (encrypted) query result. Their design allows the user to determine the sign of the query result (which indicates whether the database molecule is similar to the query, above the specified threshold), but does not let them get the actual similarity score. They show that as the number of dummies is increased, the chances of database information leaking out tends towards zero.
Of course, one could argue that the limited usage of proprietary chemical information (in terms of people who have it and people who can make use of it) means that the efforts put in to obfuscation, cryptography etc. could simply be replaced by legal contracts. Certainly, a simple way to address the scenario discussed here (and noted by the authors) is to download the remote database locally. of course this is not feasible if the remote database is meant to stay private (e.g., a competitors structure database).
But nonetheless, methods that rigorously guarantee privacy of chemical information are interesting from an algorithmic standpoint. Even though Shimuzu et al described a very simplistic situation (though the more realistic scenario where the similar database molecules are returned would obviously negate constraint 2 above), it looks like a step forward in terms of applying formal cryptanalysis to chemical problems and supporting truly safe exchange of chemical information.
Recently I came across a paper from Marth et al that described a method based on network analysis to support retrosynthetic planning, particularly for complex natural products. I’m no synthetic chemist so I can’t comment on the relevance or importance of the targets or the significance of the proposed approach to planning a synthetic route. What caught my eye was the claim that
This work validates the utility of network analysis as a starting point for identifying strategies for the syntheses of architecturally complex secondary metabolites.
I was a little disappointed (hey, a Nature publication sets certain expectations!) that the network analysis was fundamentally walking the molecular graph to identify a certain type of ring, termed the maximally bridging ring. The algorithm is described in the SI and the authors make it available
as an online tool. Unfortunately they didn’t provide any source code for their algorithm, which was a bit irritating, given that the algorithm is a key component of the paper.
I put together an implementation using the CDK (1.5.12), available in a Github repo. It’s a quick hack, using the parameters specified in the paper, and hasn’t been extensively tested. However it seems to give the correct result for the first few test cases in the SI.
The tool will print out the hash code of the rings recognized as maximally bridging and also generate an SVG depiction with the first such ring highlighted in red, such as shown alongside. You can build a self-contained version of the tool as
git clone email@example.com:rajarshi/maxbridgerings.git
mvn clean package
The tool can then be run (with the depiction output to
java -jar target/MaximallyBridgingRings-1.0-jar-with-dependencies.jar \
Recently I came across a NIPS2015 paper from Vartak et al that describes a system (APIs + visual frontend) to support the iterative model building process. The problem they are addressing is common one in most machine learning settings – building multiple models (different type) using various features and identifying one or more optimal models to take into production. As they correctly point out, most tools such as scikit-learn, SparkML, etc. focus on providing methods and interfaces to build a single model – it’s up to the user to manage the multiple models, keep track of their performance metrics.
My first reaction was, “why?”. Part of this stems from my use of the R environment which allows me to build up a infrastructure for building multiple models (e.g., caret, e1701), storing them (
list of model objects, RData binary files or even pmml) and subsequent comparison and summarization. Naturally, this is specific to me (so not really a general solution) and essentially a series of R commands – I don’t have the ability to monitor model building progress and simultaneously inspect models that have been built.
In that sense, the authors statement,
For the most part, exploration of models takes place sequentially. Data scientists must write (repeated) imperative code to explore models one after another.
is correct (though maybe, said data scientists should improve their programming skills?). So a system that allows one to organize an exploration of model space could be useful. However, other statements such as
- Without a history of previously trained models, each model must be trained from scratch, wasting computation and increasing response times.
- In the absence of history, comparison between models becomes challenging and requires the data scientist to write special-purpose code.
seem to be relevant only when you are not already working in an environment that supports model objects and their associated infrastructure. In my workflow, the
list of model object represents my history!
Their proposed system called SHERLOCK is built on top of SparkML and provides an API to model building, a database for model storage and a graphical interface (see above) to all of this. While I’m not convinced of some of the design goals (e.g., training variations of models based on previously trained models (e.g., with different feature sets) – wouldn’t you need to retrain the model from scratch if you choose a new feature set?), it’ll be interesting to see how this develops. Certainly, the UI will be key – since it’s pretty easy to generate a multitude of models with a multitude of features, but in the end a human needs to make sense of it.
On a separate note, this sounds like something RStudio could take a stab at.
As part of a project I was wondering about reports of surveys that collected chemists assessments of differnt things. More specifically, I wasn’t looking for crowd-sourcing efforts for data curation (such as the in the Spectral Game) or data collection. Rather, I was interested in reports where somebody asked a group of chemists what they thought of some particular molecular “feature”. Here, “feature” is pretty broadly defined and could range from the quality of a probe molecule to whether a molecule is complex or not.
- Brown et al, J. Med. Chem., 2015, “Understanding Our Love Affair with p-Chlorophenyl: Present Day Implications from Historical Biases of Reagent Selection“
- Li et al, Org. Biomol. Chem, 2015, “Current complexity: a tool for assessing the complexity of organic molecules“
- Franco et al, J. Cheminf., 2014, “The use of 2D fingerprint methods to support the assessment of structural similarity in orphan drug legislation“
- Sheridan et al, JCIM, 2014, “Modeling a Crowdsourced Definition of Molecular Complexity“
- Litterman et al, JCIM, 2014, “Computational Prediction and Validation of an Expert’s Evaluation of Chemical Probes“
- Kutchukian et al, PLoS ONE, “Inside the Mind of a Medicinal Chemist: The Role of Human Bias in Compound Prioritization during Drug Discovery”
- Hack et al, JCIM, 2011, “Library Enhancement through the Wisdom of Crowds“
- Oprea et al, Nat. Chem. Biol., 2009, “A crowdsourcing evaluation of the NIH chemical probes“
- Ertl et al, J. Cheminf., 2009, “Estimation of synthetic accessibility score of drug-like molecules based on molecular complexity and fragment contributions“
- Boda et al, JCAMD, 2007, “Structure and reaction based evaluation of synthetic accessibility”
- Lajiness et al, J. Med. Chem, 2004, “Assessment of the Consistency of Medicinal Chemists in Reviewing Sets of Compounds“
- Takoaka et al, JCICS, 2003, “Development of a Method for Evaluating Drug-Likeness and Ease of Synthesis Using a Data Set in Which Compounds Are Assigned Scores Based on Chemists’ Intuition”
The number of people surveyed across these studies ranges from less than 10 to more than 300. Recently there appears to be a trend towards developing predictive models based on the results of such surveys. Also, molecular complexity seems pretty popular. Modeling opinion is always a tricky thing, though in my mind some aspects (e.g., complexity, diversity) lend themselves to more robust models than others (e.g., quality of a probe).
If there are other examples of such surveys in chemistry, I’d appreciate any pointers
I have a post-doc opening in the Informatics group at NCATS, to work on computational aspects of high throughput combination screening – topics will include predicting drug combination response, visualizing large combination screens (> 5000 combinations) and so on. The NCATS combination screening platform thas tested more than 65,000 compound combinations (in checkerboard style which means more than 4.5M individual dose combinations) along with single agent dose responses. You can view publicly released data at https://tripod.nih.gov/matrix-client.
The NCATS Informatics group is a collection of very smart people, with wide ranging interests in molecular informatics. We work closely with colleagues in biology and chemistry. As a result, we eat a lot of our own dog food. In addition, we’re committed to implementing our ideas in publicly available software tools as well as publishing in journals.
Lots of data, great people and tough problems. If this piques your interest visit the job posting for more details