Assembly of microarrays for genome-wide measurement of DNA copy number.

We have assembled arrays of approximately 2,400 BAC clones for measurement of DNA copy number across the human genome. The arrays provide precise measurement (s.d. of log2 ratios=0.05-0.10) in cell lines and clinical material, so that we can reliably detect and quantify high-level amplifications and single-copy alterations in diploid, polyploid and heterogeneous backgrounds.

Quantitative analysis of chromosomal CGH in human breast tumors associates copy number abnormalities with p53 status and patient survival.

We present a general method for rigorously identifying correlations between variations in large-scale molecular profiles and outcomes and apply it to chromosomal comparative genomic hybridization data from a set of 52 breast tumors. We identify two loci where copy number abnormalities are correlated with poor survival outcome (gain at 8q24 and loss at 9q13). We also identify a relationship between abnormalities at two loci and the mutational status of p53. Gain at 8q24 and loss at 5q15-5q21 are linked with mutant p53. The 9q and 5q losses suggest the possibility of gene products involved in breast cancer progression. The analytical techniques are general and also are applicable to the analysis of array-based expression data.

Morphological similarity: a 3D molecular similarity method correlated with protein-ligand recognition.

Recognition of small molecules by proteins depends on three-dimensional molecular surface complementarity. However, the dominant techniques for analyzing the similarity of small molecules are based on two-dimensional chemical structure, with such techniques often outperforming three-dimensional techniques in side-by-side comparisons of correlation to biological activity. This paper introduces a new molecular similarity method, termed morphological similarity (MS), that addresses the apparent paradox. Two sets of molecule pairs are identified from a set of ligands whose protein-bound states are known crystallographically. Pairs that bind the same protein sites form the first set, and pairs that bind different sites from the second. MS is shown to separate the two sets significantly better than a benchmark 2D similarity technique. Further, MS agrees with crystallographic observation of bound ligand states, independent of information about bound states. MS is efficient to compute and can be practically applied to large libraries of compounds.

Molecular hashkeys: a novel method for molecular characterization and its application for predicting important pharmaceutical properties of molecules.

We define a novel numerical molecular representation, called the molecular hashkey, that captures sufficient information about a molecule to predict pharmaceutically interesting properties directly from three-dimensional molecular structure. The molecular hashkey represents molecular surface properties as a linear array of pairwise surface-based comparisons of the target molecule against a common 'basis-set' of molecules. Hashkey-measured molecular similarity correlates well with direct methods of measuring molecular surface similarity. Using a simple machine-learning technique with the molecular hashkeys, we show that it is possible to accurately predict the octanol-water partition coefficient, log P. Using more sophisticated learning techniques, we show that an accurate model of intestinal absorption for a set of drugs can be constructed using the same hashkeys used in the aforementioned experiments. Once a set of molecular hashkeys is calculated, its use in the training and testing of property-based models is very fast. Further, the required amount of data for model construction is very small. Neural network-based hashkey models trained on data sets as small as 30 molecules yield statistically significant prediction of molecular properties. The lack of a requirement for large data sets lends itself well to the prediction of pharmaceutically relevant molecular parameters for which data generation is expensive and slow. Molecular hashkeys coupled with machine-learning techniques can yield models that predict key pharmacological aspects of biologically important molecules and should therefore be important in the design of effective therapeutics.

IcePick: a flexible surface-based system for molecular diversity.

IcePick is a system for computationally selecting diverse sets of molecules. It computes the dissimilarity of the surface-accessible features of two molecules, taking into account conformational flexibility. Then, the intrinsic diversity of an entire set of molecules is calculated from a spanning tree over the pairwise dissimilarities. IcePick's dissimilarity measure is compared against traditional 2D topological approaches, and the spanning tree diversity measure is compared against commonly used variance techniques. The method has proven easy to implement and is fast enough to be used in selection of reactants for numerous production-sized combinatorial libraries.

Automatic identification and representation of protein binding sites for molecular docking.

Molecular docking is a popular way to screen for novel drug compounds. The method involves aligning small molecules to a protein structure and estimating their binding affinity. To do this rapidly for tens of thousands of molecules requires an effective representation of the binding region of the target protein. This paper presents an algorithm for representing a protein's binding site in a way that is specifically suited to molecular docking applications. Initially the protein's surface is coated with a collection of molecular fragments that could potentially interact with the protein. Each fragment, or probe, serves as a potential alignment point for atoms in a ligand, and is scored to represent that probe's affinity for the protein. Probes are then clustered by accumulating their affinities, where high affinity clusters are identified as being the "stickiest" portions of the protein surface. The stickiest cluster is used as a computational binding "pocket" for docking. This method of site identification was tested on a number of ligand-protein complexes; in each case the pocket constructed by the algorithm coincided with the known ligand binding site. Successful docking experiments demonstrated the effectiveness of the probe representation.

Scoring noncovalent protein-ligand interactions: a continuous differentiable function tuned to compute binding affinities.

Exploitation of protein structures for potential drug leads by molecular docking is critically dependent on methods for scoring putative protein-ligand interactions. An ideal function for scoring must exhibit predictive accuracy and high computational speed, and must be tolerant of variations in the relative protein-ligand molecular alignment and conformation. This paper describes the development of an empirically derived scoring function, based on the binding affinities of protein-ligand complexes coupled with their crystallographically determined structures. The function's primary terms involve hydrophobic and polar complementarity, with additional terms for entropic and solvation effects. The issue of alignment/conformation dependence was solved by constructing a continuous differentiable nonlinear function with the requirement that maxima in ligand conformation/alignment space corresponded closely to crystallographically determined structures. The expected error in the predicted affinity based on cross-validation was 1.0 log unit. The function is sufficiently fast and accurate to serve as the objective function of a molecular-docking search engine. The function is particularly well suited to the docking problem, since it has spatially narrow maxima that are broadly accessible via gradient descent.

Hammerhead: fast, fully automated docking of flexible ligands to protein binding sites.

BACKGROUND: Molecular docking seeks to predict the geometry and affinity of the binding of a small molecule to a given protein of known structure. Rigid docking has long been used to screen databases of small molecules, because docking techniques that account for ligand flexibility have either been too slow or have required significant human intervention. Here we describe a docking algorithm, Hammerhead, which is a fast, automated tool to screen for the binding of flexible molecules to protein binding sites. RESULTS: We used Hammerhead to successfully dock a variety of positive control ligands into their cognate proteins. The empirically tuned scoring function of the algorithm predicted binding affinities within 1.3 log units of the known affinities for these ligands. Conformations and alignments close to those determined crystallographically received the highest scores. We screened 80 000 compounds for binding to streptavidin, and biotin was predicted as the top-scoring ligand, with other known ligands included among the highest-scoring dockings. The screen ran in a few days on commonly available hardware. CONCLUSIONS: Hammerhead is suitable for screening large databases of flexible molecules for binding to a protein of known structure. It correctly docks a variety of known flexible ligands, and it spends an average of only a few seconds on each compound during a screen. The approach is completely automated, from the elucidation of protein binding sites, through the docking of molecules, to the final selection of compounds for assay.

Quantitative binding site model generation: compass applied to multiple chemotypes targeting the 5-HT1A receptor.

We present enhancements to the Compass algorithm that automatically deduce interchemotype relationships and generate predictive quantitative models of receptor binding based solely on structure-activity data. We applied the technique to a series of compounds assayed for 5-HT1A binding. A model was constructed from 20 compounds of two chemotypes and used to predict the affinities and bioactive conformation of 35 new compounds, most of which had new underlying scaffolds and/or functional groups. The model's mean error of prediction was 0.5 log units (essentially the assay resolution), even on quite divergent series. The predictions are supported by an interpretable hypothesis for the binding determinants of the receptor and the geometric relationships of the chemotypes.

A shape-based machine learning tool for drug design.

Building predictive models for iterative drug design in the absence of a known target protein structure is an important challenge. We present a novel technique, Compass, that removes a major obstacle to accurate prediction by automatically selecting conformations and alignments of molecules without the benefit of a characterized active site. The technique combines explicit representation of molecular shape with neural network learning methods to produce highly predictive models, even across chemically distinct classes of molecules. We apply the method to predicting human perception of musk odor and show how the resulting models can provide graphical guidance for chemical modifications.

Compass: predicting biological activities from molecular surface properties. Performance comparisons on a steroid benchmark.

We describe a new method, Compass, for predicting the biological activities of molecules based on the activities and three-dimensional structures of other molecules. The method improves on previous techniques by representing only the surface of molecules, by incorporating a nonlinear statistical method, and by automatically choosing conformations and alignments of molecules. We use a benchmark problem of steroid binding affinity prediction to compare the performance of the method with that of two previous systems: CoMFA and a molecular similarity method. Compass predicts steroid affinities substantially more accurately than the others, which represent the state of the art. We present experiments showing that the improved performance depends on each of the technical innovations.

Pitfalls in the lateral imaging of the thyroid with pinhole collimators.

Imaging of the thyroid gland with pinhole collimators in the lateral projection is associated with the problem of count rate contribution from the contralateral lobe. In a series of experiments conducted using a thyroid-simulating phantom, the contribution of this nature was calculated. Radioisotopes of different energies such as 99mTc and 131I were used to see the effect of energy on this contribution. It was observed that the count rate contribution from the contralateral lobe to the lobe proximal to the collimator can be as high as 70% at a distance of 6 cm from the collimator. At distances more than 6 cm these contributions from the contralateral lobe are much more than this.

Quality control of dose calibrators.

Radionuclide dose calibrators are instruments that require regular checks and proper usage to ensure that they are operating correctly. The responsibility for the accuracy of the activity and the resultant estimate of the dose delivered to the patient rests with the persons administering the radiopharmaceutical. Hence, it is essential that accurate measurements of radioactivity be provided by the dose calibrator in use in every nuclear medicine department. The routine testing of the calibrator would ensure the user that the overall characteristics of the instrument are within acceptable limits. The quality control of two dose calibrators was carried out elaborately and their performance was evaluated from various aspects of quality control such as geometry effect, instrument linearity, precision test, instrument accuracy etc. It was found that one of the calibrators was having non-linear response resulting in an error of 18% from the decay-predicted activity. It also showed a saturation effect beyond 310 mCi. Hence, it needed repair or recalibration. However, the other calibrator was working satisfactorily. This study also helped us in realizing certain operational pitfalls of the dose calibrators.

Scan appearances in hydatid cysts of the liver: analysis of 55 cases.

Liver scans done at our Centre between 1964 and 1977 were analyzed with reference to hydatid cysts in liver. Among the approximately 7200 patients analyzed, there were 55 suspected to have hydatid cysts. The criteria adopted in screening the patients for hydatid cysts were based upon clinical findings and liver scan findings. The cases were analyzed with respect to their histologic findings, locations of cold areas in the liver, and correlation with Casoni's test. We found that 1) the cystic involvement was most often in the right lobe of the liver; 2) Casoni's test does not rule out or establish the presence of hydatidosis; 3) radioisotopic vascular studies done in some cases contributed to the diagnosis; 4) although a positive liver scan does not indicate the cause of the disease, the scans showing large and totally photon-deficient cold areas along with good general condition of the patients help in the diagnosis of hydatid cysts in the liver.

Efficacy of brain scanning in epilepsy of late onset.

Brain scans of 513 patients with epilepsy of late onset were analysed with reference to the patient's age and sex and to the nature of convulsion. Only 17 of them showed an abnormal concentration of radionuclide indicating a space-occupying lesion in the brain. The findings of those patients who had positive brain scans were correlated with EEG findings. It was found that the incidence of epilepsy of late onset is almost 3 times higher in males than in females and that the age cannot be considered as a criterion for screening the patients for brain scan investigations as far as epilepsy of late onset is concerned. In our opinion, the incidence of 3.3% is not too low. A positive brain scan finding calls for further investigation and helps in deciding the management and further line of treatment of the patients. Moreover, a normal scan rules out the presence of a space-occupying lesion and helps as a screening procedure.