Pros and cons of alignment-independent descriptors

Working with molecules in 3D is computationally expensive compared to most 2D methods. Most modern cheminformatics toolkits can do hundreds of thousands to millions of 2D fingerprint comparisons per second, with 3D similarity techniques being multiple orders of magnitude slower.

The computationally-expensive part of the 3D calculations usually involves aligning conformations to each other. The natural tendency therefore, is to see if we can skip this step and compute a set of properties that can tell us if two molecules are similar in 3D without actually having to align them. If this works we can get the best of both worlds: the speed of 2D comparisons combined with the accuracy and structural independence of 3D similarity functions.

Pharmacophoric descriptors

The earliest version of this idea is the simple pharmacophore. All you have to do is assign a few pharmacophoric points to each molecule (usually based on some sort of functional group pattern recognition), then generate a set of descriptors based on sets of these (usually 2 or 3). If two molecules share one or more pharmacophore descriptor, then they match.

Pharmacophore searches succeed on some counts: they are indeed very fast, and they do encode some 3D information. However, they involve a very crude binning of the wide array of possible intermolecular interactions into a few pharmacophore types, and they describe shape poorly giving them very limited predictive power.

If pharmacophores can’t describe shape well, are there other techniques that can? A number of different methods have been presented, such as those based on multipole moments or spherical harmonic coefficients (e.g. ParaSurf/ParaFit), as well as methods based on statistical moments such as Ultrafast Shape Recognition (USR). None of these has achieved widespread use: harmonic coefficients are not rotation-invariant, while the USR technique correlates poorly to more accurate measures of shape similarity. 1

Field descriptor distances

It would be nice if there was a way of providing alignment-independent descriptors which described both electrostatics and shape/pharmacophoric properties with a reasonable degree of accuracy. This is actually one of the first things we did when we were looking into starting Cresset – we developed a method called FieldPrint that encodes the distance matrix of field descriptors down into a fingerprint that can be used for alignment-independent similarity calculations. The concept is similar to that of GRIND 2 which was published around the same time, although the algorithmic details are somewhat different.

We put a lot of work into these techniques, but were never able to get a method that we were completely satisfied with. The problem we found is that encoding the distances in pairs/triplets of field descriptors ends up losing too much 3D information, and as a result you either end up with a slower mimic of standard 2D fingerprints, or you end up with a large false positive count. The FieldPrints have a tendency to find molecules with a similar overall pattern of positive/negative field, but can compute a very high similarity for molecules that are in reality quite dissimilar in terms of the 3D spatial arrangement of those fields. My belief now is that this is an inherent flaw of alignment-independent descriptors: they either have to be sufficiently complex that you are in effect computing an alignment, or you lose too much information and are not significantly better off than just using old-fashioned structural/pharmacophoric fingerprints.

As you move from full 3D interaction potentials to 2D correlograms to 1D fingerprints comparisons get faster but you lose information
Figure 1: As you move from full 3D interaction potentials to 2D correlograms to 1D fingerprints, comparisons get faster but you lose information

Handling conformation

A further consideration is how you handle conformation. The original GRIND papers just use a single conformer per molecule, and their validation was confined to series of rigid molecules or sets of molecules where single conformations were generated and manually adjusted to be similar. In the general case neither of these shortcuts will work. Any method that purports to be 3D but starts with a single conformation per molecule is inherently flawed: the whole point of 3D is that molecules are flexible.

There is a disturbing number of papers out there that do some sort of notionally-3D analysis on set of single CORINA-derived conformations. You can get very good enrichment factors on retrospective virtual screens doing this, but in practise the enrichments are largely bogus. CORINA is deterministic, and as a result molecules with similar structures will tend to be put into similar conformations. Combine this with the fact that many standard retrospective VS data sets have very low structural diversity, and the problem becomes apparent. The query molecule and its dozens of congeners in the “actives” data set are all placed in the same single conformation, and so application of a 3D or pseudo-3D technique can easily produce excellent-looking enrichment statistics. However, the enrichment all comes from a hidden 2D similarity.

So, single-conformation methods are a dead end and we need to consider flexibility. Once you are doing so, you need to factor in both the conformer generation time as part of the build time for the descriptor, and also factor in that your comparison speeds will now be two to four orders of magnitude slower than 2D fingerprints (assuming 100 conformers per molecule, and depending on whether you know a single bioactive conformation for one of the two molecules being compared or whether you need to compare conformer populations). 2D methods thus have an unassailable speed advantage, which is part of the reason they remain so popular.

Using FieldPrint as a filter

Our original vision for Blaze (or FieldScreen, as it was then) was that it would rely on the FieldPrints to give extremely rapid searching. You can get quite good enrichment factors from the FieldPrints in retrospective virtual screens, but when we investigated further this is largely because they act as a proxy for overall molecular size and charge. Once you control for that by more careful selection of decoys the FieldPrint performance is much less good. Analysing a molecular similarity technique through retrospective virtual screening performance is very very hard to do well, and as a result I am intrinsically wary of methods that present a set of DUD enrichments as their sole validation: FieldPrints perform quite well on DUD, but we know that they are not particularly effective in real prospective applications.

We still use the FieldPrint technology: it’s the first search stage in every Blaze run. It’s generally good enough to filter out 25-50% of decoy molecules that have no similarity to the query, but certainly not good enough to use the FieldPrint ranks directly. This is why we just use them as a pre-filter: molecules that pass that filter have much more accurate similarities computed using our alignment-based clique/simplex algorithms.

In the end, there’s no real short cut. All attempts to date to make 3D comparisons faster by simplifying descriptors and skipping the expensive alignment step just seem to leave out too much information – such techniques can be useful for cutting down the search space but if you’re going to spend CPU cycles working in 3D you might as well do it properly!

1. T. Zhou et al. / Journal of Molecular Graphics and Modelling 29 (2010) 443–449
2. Pastor, M.; Cruciani, G.; McLay, I.; Pickett, S.; Clementi, S. J. Med. Chem. 2000, 43 (17), 3233–3243.

Download an evaluation

Try our software for yourself – download a free evaluation.

Understanding SAR of circadian rhythm modulators

Anupriya KumarAnupriya Kumar, Assistant Professor, Nagoya University, summarizes the publication Development of Small-Molecule Cryptochrome Stabilizer Derivatives as Modulators of the Circadian Clock which cites how Cresset software was used to understand SAR of circadian rhythm modulators.


Mammalian biological clocks control almost every physiological process such as metabolism, hormone secretion, body temperature and the sleep/wake cycle. Each cell in our body follows a circadian rhythm of 24 hours. Our modern lifestyle can disrupt this rhythm, for example working night shifts, or flying long distance. This in turn can lead to sleep and metabolic disorders such as obesity and type 2 diabetes. The daily oscillations of the circadian rhythm are generated by a loop which consists of clock genes such as CLOCK and BMAL1 (transcriptional activators) and two repressors: Cryptochrome (CRY1 and CRY2) and Period (PER1 and PER2).

Motivation to study the circadian rhythm modulators

The circadian rhythm can be modulated by small molecules, for example carbazole derivatives. One such molecule, KL001 was discovered by screening 60000 structurally diverse compounds in a cell-based Bmal1-dLuc reporter assay 1. KL001 lengthens the circadian rhythm by nearly 6 hours by stabilizing the clock protein cryptochrome (CRY). In the co-crystal structure of KL001 with the CRY2 protein (PDB ID. 4MLP 2), KL001 displaces the naturally binding FAD ligand (flavin adenine dinucleotide) while occupying only a small part of the binding pocket. Subsequent derivatization of KL001 resulted in a ten times more potent compound, KL044.

Building 3D-QSAR model using Forge

To understand the factors responsible for binding and stabilizing the CRY protein, we used Forge to perform 3D modeling and multivariate statistical analysis on 60 derivatives of KL001 and KL044. For generating a pharmacophore based on KL001 and KL044, we used FieldTemplater, which could reproduce the binding pose of KL001 to the CRY protein as shown in the crystal structure. We established a robust 3D-QSAR model by aligning all 60 compounds to the pharmacophore. The molecules were randomly divided into a training set of 54 and a test set of 6 compounds. Using this model, we successfully predicted activities of 5 new compounds and explained the evolution of the SAR by analyzing the contributing field points. By overlaying the field points from the 3D-QSAR model with the protein, we deciphered the important ligand-protein interactions.

Activity Cliff analysis

Activity Miner was used to study differences in the activity of the compounds by analyzing the electrostatic potential. The differences in the electrostatic potential maps (red = positive potential, cyan = negative potential) of the two matched pairs are shown below. For example, introducing a disubstituted phenyl ring and an amide group (a) enhances the activity as does the presence of a mesyl group in the same position (b).

Activity Cliff analysis


The current study serves as an encouragement for interdisciplinary teams to collaborate by showing that theoretical hypothesis and experimental studies complement each other. Knowledge based functional modifications can be made to KL001 and KL044 based on the activity enhancing key field points obtained from the 3D-QSAR model. Based on these findings, virtual screenings of compound databases can be performed to find novel compounds which would help in forwarding the field of drug discovery for circadian rhythm modulators which could lead to potential drugs for diabetes and other metabolic disorders.


1. Identification of small molecule activators of cryptochrome. Hirota T, Lee JW, St John PC, Sawa M, Iwaisako K, Noguchi T, Pongsawakul PY, Sonntag T, Welsh DK, Brenner DA, Doyle FJ 3rd, Schultz PG, Kay SA. Science. 2012 Aug 31;337(6098):1094-7.
2. Nangle S, Xing W, Zheng N. Cell Res. 2013 Dec;23(12):1417-9.

Affordable virtual screening with Blaze: Benchmarks


We released BlazeGPU a couple of years ago, allowing the full power of the Blaze virtual screening system to be used on a few consumer graphics cards rather than a full-scale Linux cluster. Since then, graphics cards and CPUs have only got faster, so we decided that it was time to update our benchmarks and see how well all of the new hardware performs.

For these benchmarks we took a random subset of 4,000 molecules from our in-house Blaze data set and searched with a medium-sized query molecule. The molecules in the data set average 80 conformers each. We’ve run with three different search conditions: the full slow-but-accurate simplex algorithm, the standard clique algorithm and the new fastclique algorithm. All of these were run with 50% fields and 50% shape.

CPU performance

Firstly, the CPU benchmarks. All of these are single-core performance, but with all cores loaded so that we’re not benefitting from Intel Turbo Boost. In most cases Blaze will be saturating all cores, so this is representative of real-world performance. Note that the vertical axis is on a log scale.

CPU benchmarks

As can be seen, there’s a significant performance difference between the older CPUs at Cresset (such as the Q6600) and the newer Ivy Bridge i7-3770K chips, but not nearly as much as you would expect given that the Q6600s are around 7-8 years old at this point. The significant speed improvements of the fastclique algorithm are clearly visible with the throughput being more than 4x greater than the original clique algorithm. The last set of columns on the graph are from an Amazon c4.xlarge instance and show that the performance of each core on those systems is roughly the same as the Sandy Bridge i3-2120.

GPU Performance

Moving on to the GPUs, we’ve tested the throughput on a variety of different systems. Firstly, we’ve tested a variety of GTX580s on different motherboards and processors. As you would expect, for the most part the performance is governed by the GPU, but the exception is the fifth test system which is noticeably slower than the others. That card is sitting in a much older chassis with an older motherboard and hence is probably suffering from lack of backplane bandwidth to the GPU.

GPU benchmarks

The newer GTX960s perform extremely well on the Blaze calculations. We weren’t sure if they would, after the disappointment of the GTX680 which was noticeably slower than the 580 (data not shown). The difference is noticeable in the clique stages, but really stands out in the simplex calculations where a GTX960 is 50% faster than the GTX580s. By contrast, the high-end Tesla hardware is not a great performer on the Blaze OpenCL kernels. By all accounts the Tesla hardware is significantly faster than the consumer hardware on double precision workloads, but the Blaze code is all single precision and in that realm the cheap consumer hardware has an unbeatable price/performance advantage.
Finally, the GRID K520 is the hardware found on the Amazon g2.2xlarge and g2.8xlarge instances. As can be seen, it’s not a brilliant performer on the Blaze workload, being around the same speed as the Tesla on the fastclique algorithm but noticeably slower than all of the other cards tested on the simplex workload. However, it provides a nice test of GPU scaling: when running on a 4 times larger data set on all 4 GPUs of a g2.8xlarge instance, we observed substantially the same throughput as running the original data set on a single K520 GPU, showing that we can parallelise across multi-GPU systems with no loss of performance.

Cost efficiency on Amazon

Converting the throughput shown above, we can look at the cost of screening on the Amazon cluster with Blaze. The raw cost to screen a million molecules is shown in the table. Note that the actual costs will be somewhat higher, due to job overheads and data transfer costs.

Cost efficiency on Amazon

The Amazon GPU solutions are noticeably cheaper for fastclique jobs, roughly cost-competitive for the clique runs, but the poor performance of the K520 on the simplex task means that it is significantly more expensive there. As a result, at the moment there’s no real impetus to use the Amazon GPU resources unless you can get them significantly more discounted than the CPU instances on the spot market.


New hardware is significantly faster at running Blaze than old stock as would be expected. However, the speed increases are much lower than they have been in the past, with CPUs that are well past their best still performing adequately. On the GPU side, Blaze performs particularly well on commodity graphics cards leaving few reasons for us to invest in dedicated GPU co-processing cards.

The cost of running a million molecule virtual screen on the Amazon cloud has never been cheaper. If tiered processing is used as is the default for Blaze then these screens can be performed for a very low cost indeed – less than $15 per million molecules for the processing costs.

Contact us for a free evaluation to try Blaze on your own cluster, or Blaze Cloud.

Bioisosterism – harder than you might think

A common concept in medicinal chemistry is the bioisostere. The exact definition of a bioisostere is rather fuzzy, but the Wikipedia entry is as good as any: ‘bioisosteres are chemical substituents or groups with similar physical or chemical properties which produce broadly similar biological properties to another chemical compound.’ Bioisosteric replacements come in two general classes: core replacements, where the centre of a molecule is changed, creating a new chemical series; and leaf replacements where a replacement is sought for part of the molecule on the periphery, keeping it in the same series. Core replacements can get you out of a difficult ADMET or IP situation or can be a useful stepping stone to developing a backup series. Leaf replacements are more common as part of the day-to-day work of lead optimization – having discovered a substituent that works, the obvious thing to do is to try more things like it!

There are many software products available to search for bioisosteric replacements, both commercial and freely-available. The proliferation of methods is largely because searching for bioisosteres looks so easy. Look at the piece you want to replace, search a database for something largely the same size and geometry, and present the results. Simple, right? However, if you delve into it, the whole process is much more complex than it appears …

The first issue for a bioisostere replacement method involves deciding whether a fragment is physically the right size – are the connection points in the right place and at the right angle? Again this seems a trivially easy question to address until we consider that this is a 3D question. You need to consider the 3D geometry of the fragment – what conformation is it in? We chose to use a mixed approach to this: we provide both fragments derived from the CSD where the conformation comes directly from the small molecule crystal structure and large databases of fragments where we provide a small distribution of low energy conformations. As a result, you can just search experimental small molecule crystal conformations if you wish, but you can augment that search with a much larger and richer conformation database if you so desire.

The second criterion is pharmacophoric: a bioisosteric replacement must have ‘broadly similar biological properties’ and hence its interactions with the protein must be similar to the original molecule. Again, there’s a subtlety here. If I’m looking for a replacement for a triazolothiazole, then when I’m presented with a candidate fragment the obvious thing to do is ask “how similar is it to triazolothiazole?” This approach is flawed because the properties of the candidate fragment (and the initial triazolothiazole!) depend on their environment i.e. the rest of the molecule. This is especially true if you are interested in molecular electrostatics, as the electrostatic potential of a molecule is a global property that cannot generally be piecewise decomposed. That is why Spark performs all scoring in product space, not in fragment space (Figure 1). However, even if you are using a more crude method such as shape or pharmacophore similarity, treating the fragment in isolation can be very misleading.

Figure 1: Merging a new fragment can (subtly) change the electrostatics of the rest of the molecule. Red=positive, Blue=negative regions.

Working in product space brings advantages in steric considerations as well as electronic. Even though a particular fragment might have a beautiful geometric fit, and match the original core very well in terms of shape/electrostatics/color/pharmacophores/whatever, you still need to assess whether the fragment is compatible with the conformation of the original molecule. That assessment cannot be done in fragment space! In Spark we handle this in two ways. Firstly, all product molecules are minimized prior to scoring. If, due to steric or electronic effects, the candidate fragment just is not compatible with the desired conformation of the original molecule, then this minimization will distort the invariant parts and lead to a very low similarity score. Fragments which cause steric clashes or lone pair-lone pair repulsions are thus automatically filtered out (Figure 2). Secondly, we follow this up with an explicit assessment of strain energy around the newly-formed bonds, so the user can immediately see whether there are any causes for concern.

Toluyl is similar to phenyl except when it causes a steric clash
Figure 2: Toluyl is similar to phenyl, except when it causes a steric clash. Whether it does or not depends on what R is!

Another thing that can be drastically affected by the environment is the hybridization and charge state of the fragment (and the rest of the molecule). The main culprit here is nitrogen. Is an amine basic or not? Is the nitrogen pyramidal or trigonal planar? Both these questions depend strongly on what the nitrogen is attached to and hence you cannot assess the suitability of a fragment without reference to the chemical environment that you are going to place it in. These questions turn out to be surprisingly difficult to handle in a completely robust way: there is a large amount of code in Spark which assesses the local chemical environment around the newly-formed bonds and determines whether any hybridization or formal charge changes are required. It turns out you cannot just recharge the product molecule, as in some cases the user may have assigned particular charge states to parts of the molecule (based on experimental knowledge of the charge state when bound to the protein, for example) and you don’t want to undo that.

The final complication to performing the scoring properly (i.e. in product space) is accounting for the available degrees of freedom. If you are doing a core replacement, then the positon of the core is generally completely determined by the attachment points. However, if you are doing a leaf replacement, then the question arises as to what rotamer to choose around the attachment bond. The only solution is to perform a limited conformation search around that bond and score each conformer, which has the potential to get very slow although there are some computational tricks you can do to reduce the search space. It’s not just leaf groups that have to be searched: replacing the centre of a molecule might require a rotation scan if there are two attachment points which happen to be collinear.

So, our nice and simple bioisosterism calculation has become rather more complicated. Rather than just see if a fragment fits into the hole in the original molecule, we need to merge it in properly, assess any hybridization or formal charge state changes that are required, minimize the resulting molecule, perform a rotation scan around any underspecified degrees of freedom, recompute the electrostatic potentials, align to the original molecule and calculate the steric and electrostatic similarities. Only then can we decide whether the fragment is any good or not! Luckily, Spark does all of that for you, so as a user you can concentrate on the results, rather than on the complicated calculations required to produce those results.

Try it for yourself – download a free evaluation of Spark.

Electrostatics of the Topliss tree

Since its publication on J. Med. Chem in 1972, J.G. Topliss’ operational scheme for aromatic substitution,1 commonly known as ‘the Topliss tree’, has been widely used by medicinal chemists to explore the substitution pattern on a benzene ring.

This decision tree approach was designed for maximum simplicity of application by the medicinal chemists, avoiding use of computers and statistical procedures. It is based on a stepwise selection of substituents designed to progress as rapidly as possible to the most potent compounds in a series.

The Topliss tree is based on the concept, pioneered by Hansch,2,3 that the observed changes in activity following the introduction of a substituent on a benzene ring depend on its lipophilic, electronic and steric properties, which in Topliss’ original publication were represented in a quantitative manner by the physico-chemical parameters µ, σ and Es.

To complement this traditional drug design approach, we have used Torch,4 Cresset’s powerful molecular design tool for medicinal and synthetic chemists, to visualize and give new insight to the deep changes in electrostatic and hydrophobic properties introduced by the substitution patterns recommended by the Topliss tree:

Electrostatics_Topliss tree

Download ‘Electrostatics of the Topliss tree’

1. J. Med Chem. 1972, 15, 1006
2. J. Amer. Chem. Soc. 1964, 86, 1616
3. ‘Drug Design’, Vol I 1971, 271

Is this compound worth making?


In deciding if a compound is worth making medicinal chemists need to consider a number of factors including:

  • does it fit with the known structure activity relationship (SAR)?
  • does it explore new interactions or regions of space?
  • does it have good physico-chemical properties?

Summarizing the SAR content of large compound series in a single informative picture is challenging. Traditional 3D-QSAR techniques have been used for this purpose, but are known to perform poorly where the structure-activity landscape is not smooth. 3D Activity cliff analysis has the potential to explore such rugged SAR landscapes: pairs of compounds with a high similarity and a large difference in activity carry important information relating to the factors required for activity. However, looking at pairs of compounds in isolation makes it hard to pinpoint the changes which consistently lead to increase potency across a series.

Similarly, summarizing the regions and interactions that have been explored in a single picture is not trivial, especially if the electrostatic character of compounds is to be considered (e.g., replacement of electron rich with electron poor rings).

Here we present Activity Atlas, a new technique to analyse a series of compounds and derive a global view of the structure-activity relationship data.



The Regions explored summary gives a comprehensive 3D picture of the regions explored in electrostatic and shape space. A novelty score is assigned to each compound, enabling to predict whether newly designed candidates are likely to contribute additional SAR knowledge, and are thus worth making.

Average of actives summarizes electrostatic and shape properties which consistently lead to potent ligands across the series.

The Activity cliff summary gives a global view at the activity cliff landscape highlighting the regions where either more positive/negative electrostatic potential, or larger steric/hydrophobic potentials increase activity.

A traditional 3D-QSAR model2 was built on the same data set (q2 = 0.7). While 3D-QSAR seems better at extracting information where SAR is continuous, Activity Atlas gives more definition in regions where SAR requirements are critical.

Activity Atlas gives more definition in regions where SAR requirements are critical

Accounting for uncertainties in the alignment

Activity Atlas takes into account multiple alignments, weighted by their probability to be correct:

Accounting for uncertainties in the alignment1
Based on the distance from the top result

Accounting for uncertainties in the alignment2
Based on the absolute similarity value (CCDC-AZ dataset)

Analyzing multiple alignments per molecule enables the technique to account for uncertainties in the orientation of flexible side chains not represented in the reference(s).

Application to selectivity

Application to selectivity

Application to a set of adenosine receptor antagonists with activities against A1, A2a and A3 receptors3 demonstrates the utility of the technique in locating critical regions for selectivity. Examination of the steric and electrostatic maps for the three subtypes clearly shows which regions should be targeted in order to enhance subtype selectivity.

In the example above, the right hand side of the molecules can be used to discriminate between A3 and the other two subtypes, while A1 and A2a can be separated by increasing steric bulk and positive charge around the top of the molecules.


The Activity Atlas technique is a powerful way of summarizing SAR data in 3D. By combining information across multiple pairs, it enables a global view of the critical points in the activity landscape. The Average of actives summary captures in one picture the 3D requirements for potency, while the Regions explored summary enables prioritizing compounds which add crucial SAR information over trivial analogues.


1. J. Chem. Inf. Model. 2006, 46, 665-676
3. J. Chem. Inf. Model. 2011, 51, 258-266

Rapid technique for new scaffold generation II: What is the best source of inspiration?


Scaffold hopping and R-group replacement remain central tasks in medicinal chemistry for generating and protecting intellectual property. Spark is a bioisostere replacement tool (available as a desktop software application) for rapidly generating reasonable yet novel scaffold and R-group replacements using Cresset’s molecular field points.

Cresset’s field technology condenses the molecular fields down to a set of points around the molecule, termed ‘field points’. Field points are the local extrema of the electrostatic, van der Waals and hydrophobic potentials of the molecule.

field points

Spark workflow

The Spark approach uses a database of molecule fragments, or available reagents, to suggest replacements that maintain the shape and electrostatic character of a known active molecule. The user identifies the region of a known active
molecule that they wish to replace, and this piece is removed.

region of known active molecule to replace
The number of bonds broken is recorded together with the distance and angle between any pair of broken bonds. This information is used to search a database of fragment conformations for replacement moieties.

search for replacement moieties
The product molecule is energy minimized and then scored as a replacement. Scoring is performed using an average of field and shape similarity on the product molecule. Scoring the product (rather than the fragment) allows the electronic changes induced in the rest of the molecule to be taken into account.

scored as a replacement
By default, the scoring reflects the change relative to the original molecule, but the user can choose to add other molecules that can be used in the scoring. In this way compounds with sub-optimal interactions can be improved by mimicking other known actives.

Fragment sources in Spark

Spark generates bioisosteres from databases of fragments derived from:

  • Commercially available, real compounds and reagents (ZINC)
  • Theoretical aromatic rings (VEHICLe)
  • Literature reports of bioactive compounds (ChEMBL)
  • Fragments from the Cambridge Structural Database (CSD) of small molecule crystal structures

In this case study we investigate which of the fragment sources available in Spark is the best source of inspiration.

best source of inspiration
If you have access to significant proprietary chemistry, to specialized reagents, or want to consider fragments from reagents that you have in stock, then the creation of custom databases with the Spark Database Generator will enable you to exploit your own proprietary chemistry to generate and protect intellectual property.

R-group replacement to D3 antagonists

The ChEMBL ‘common’, ‘rare’ and ‘very rare’, ZINC ‘very common’, ‘common’, ‘less common’, ‘rare’, ‘very rare’, ‘singleton’ and the VEHICLe fragment databases were searched using ‘Accurate But Slow’ calculation settings. Compounds with piperazine scaffolds were filtered out as these are very well known in the literature.

known scaffolds
Known D3 scaffolds were found in ChEMBL or ZINC (commercially available compounds) databases. Novel solutions were found in the ChEMBL database.

An analysis of the chemical diversity of the known D3 scaffolds retrieved from each database clearly shows that the less common fragments derived from the literature database are a precious source of potentially
useful chemical diversity. Note that these less common fragments may be associated with more complex and less documented synthetic routes.

analysis of chemical dversity

Scaffold hopping application to Sildenafil

The ChEMBL and VEHICLe fragment databases were searched using ‘Accurate But Slow’ calculation settings. The protein structure for 1UDT was used as an excluded volume, constraining the field points associated with the interaction with glutamine (Gln817) in the 1UDT protein.

known actives were found
Known actives were found in ChEMBL and VEHICLe databases. Novel but highly plausible solutions were found in the VEHICLe database.


Spark provides both known active scaffolds and novel solutions that represent opportunities for scaffold hopping and R-group replacement.

The nature of the experiment appears to dictate the best source of fragments. It is therefore important to have a wide range of fragment sources to choose from for each experiment, to provide a balance between novelty and synthetic accessibility.

The creation of fragment databases from proprietary collections of compounds can be a powerful way of increasing the chemical diversity available to Spark.


Examining the diversity of large collections of building blocks in 3D


2D fingerprint-based Tanimoto distances are widely used for clustering due the overall good balance between speed and effectiveness. However, there are significant limitations in the ability of a 2D fingerprint-based method to capture the biological similarity between molecules, especially when conformationally flexible structures are involved. Structures which appear to largely differ in functional group decoration may give rise to quite similar
steric/electrostatic properties, which are what actually determine their recognition by biological macromolecules.

In BioBlocks’ Comprehensive Fragment Library (CFL) program, we were confronted with clustering a very large collection of scaffolds generated from first principles. Due to the largely unprecedented structures in the set and our design aim to populate the 3D ‘world’, using the best 3D metrics was critical. The structural diversity of the starting collection of about 800K heterocyclic scaffolds with variable functional group decoration was not adequately captured by 2D ECFP4 fingerprint Tanimoto distances, as shown by the rather flat distribution of 2D similarity values across the set, and by their lack of correlation with the 3D similarity metrics.

The initial step of any clustering procedure is the computation of an upper triangular matrix holding similarity values between all pairs of compounds. This step becomes computationally demanding when using 3D methods, since an optimum alignment between the molecules needs be found taking into account multiple conformers.

The presentation covers the methodological and technical solutions adopted to enable 3D clustering of such a large set of compounds. Selected examples will be presented to compare the quality and the informative content of 3D vs 2D clusters.


See presentation ‘Examining the diversity of large collections of building blocks in 3D‘ given at 250th ACS national meeting.

Is it worth making? Assessing the information content of new structures


We have recently presented a method of summarizing the information obtained from 3D activity cliff analysis: examination of all pairs of molecules can distinguish between apparent cliffs that are outliers, or due to measurement error, and those which consistently point to particular electrostatic and steric features having a large impact on activity. To do this it has proved essential to allow for alignment noise: no 3D alignment technique is perfect, so we apply a Bayesian analysis to correct for potential misalignments and for the case where a molecule is aligned correctly except for a flexible substituent whose conformation is under-constrained. We use the recent AZ/CCDC alignment validation data set to determine valid estimates for the Bayesian priors.

As an extension of this technique, it is possible to mine the data for a simple picture of explored pharmacophoric space, corrected for the conformational and alignment flexibility of each molecule. This provides an invaluable picture to the chemist of which parts of property space around a molecule have been adequately explored. When considering a new molecule for synthesis, it is possible to compute the amount that this would increase the explored pharmacophoric space and hence present an ‘information content’ score for the new molecule: if we made and tested this new molecule, how much would it actually increase the structure activity relationship (SAR) information content of the data set?

The combination of this, with the activity cliff summary data, allows a simple qualitative evaluation of the SAR of a data set in 3D, alongside guidance on which parts of pharmacophoric space have been mined out and which remain underexplored. We present the application of these techniques to several literature data sets.


See presentation ‘Is it worth making? Assessing the information content of new structures‘ given at the 250th ACS National Meeting.

Scaffold hopping into new DPP-IV protease inhibitors


Cresset’s powerful scaffold hopping and fragment replacement software can be used to mimic an existing ‘active’ molecule’s electrostatic patterns to quickly, and efficiently, generate new molecular designs. The outcome, based on 3D molecular electrostatic similarity, is more biologically relevant than that from other similarity metrics. Here, we show how Spark can be used to rapidly generate potentially valuable chemical ideas for new DPP-IV inhibitors.


Two peptide hormones, GLP-1 and GIP, mediate lowering of blood glucose level through stimulation of insulin release and inhibition of Glucagon release. DPP-IV cleaves a dipeptide from the N-terminus of both GLP-1 and GIP hormones to give their inactive forms, thus abolishing their glucose lowering action. DPP-IV inhibitors have been shown to be important agents useful for treating type II diabetes.

Scaffold hopping methodology

The method involves:

  • An initial bioactive conformation: Preferably a potent target molecule of interest. This can either be extracted from an X-ray or modeled (e.g. from a pharmacophore or alignment / binding hypothesis).
  • Field pattern generation: Cresset’s proprietary XED force field used to generate electrostatic properties.
  • Database of molecule fragments: Spark uses an internal database of fragments derived from ChEMBL and ZINC. Custom sets can be generated from proprietary chemistry.
  • Automated reconstruction of the new 3D molecules: Can specify linking chemistry and replacement sites (e.g. scaffold or decoration).
  • Aligns and scores output as full minimized molecules: Using ’field similarity’ and shape similarity relative to the starting template(s).
  • Protein target can be used as excluded volume: Can take account of protein pocket by penalising examples with steric clashes.
  • Filter output on physiochemical properties.

DPP-IV X-ray ligand electrostatic and shape similarity analysis

A plethora of molecules are already known which inhibit DPP-IV and have useful anti-diabetic properties.

Hierarchical clustering of DPP-IV x-ray ligands by electrostatic and shape similarity
Figure 1. Hierarchical clustering of DPP-IV x-ray ligands by ‘electrostatic and shape similarity’.

Superimposition of over 20 published x-ray crystal structures, followed by hierarchical clustering using all-by-all field similarity on the ligands, reveals two main clusters and a distinct mode of binding for a fluoro-olefin example (Figure 1). This unique clustering, performed in Cresset’s ligand-focused workbench, Forge, allowed the selection of three distinct inhibitors for further scaffold hopping work.

Experiment and results

Alogliptin (1), Omarigliptin (2) and the fluoro-olefin (3, PDB: 3C45) represent some of the most ligand efficient examples from these clusters. Two experiments were performed in Spark using (1) and (2) in a simple scaffold hopping exercise. A final experiment was a chemotype merging experiment: a truncated (2) was used, with (3) as a second template, to find molecules bridging the two series. Workflow as shown in Figure 2.

Example results (Table 1) shows the diverse range of output suggestions provided for new chemistry and validates the method by providing examples which already have precedents in patents and the literature.

Scaffold hopping and merging_input structures fields field points and output examples
Figure 2. Scaffold hopping and merging – input structures, fields, field points and output examples.

Diverse range of output suggestions provided for new chemistry
Table 1. Scaffold hopping and merging examples of output 2D structures.


The Spark searches output a wide range of diverse chemotypes that included active or very close architectures to known active frameworks. A number of these had been discovered through HTS rather than through rational design. Tight control over the chemistry ensures that feasible chemistry is provided from known fragments whilst maintaining the features necessary for activity.


Spark is a powerful molecular modeling tool for the rapid virtual elaboration of scaffold ideas; either in scaffold hopping, merging, fragment growing or linking experiments. Applying Spark to chemotypes bound to the active site of DPP-IV provided a range of interesting and synthetically-feasible suggestions.


Green et al, Diabetes and Vascular Disease Research 2006, 3 No3, p159-65.

Protein pictures were rendered using open source Pymol from Delano Scientific.

[This content was presented as a poster at Proteinase 2015.]