PathGrid: a service-orientated architecture for microscopy image analysis.

This paper describes 'PathGrid'--an analysis and data integration system, developed initially to meet the demands in the analysis of medical microscopy imaging data. An overview of the current system is given, describing the techniques used in developing the data handling infrastructure and the analysis algorithm development. The use of software created in the context of systems designed for the astronomy domain is noted, specifically infrastructure from the astronomy virtual observatory movement for data discovery, access and workflow management, and astronomical image analysis software adapted for the analysis of high-throughput astronomy imaging surveys. This paper notes the applicability of the techniques from the astronomy domain. The testbed infrastructure deployment is described, emphasizing its speed and ease of use and support. The validity of the analysis techniques is confirmed through the pilot study described here--with the application to a large sample of immunohistochemistry microscopy data obtained in part for assessing the oestrogen receptor status of breast cancers. The analysis showed that the specificity and sensitivity values for the automatic scoring using PathGrid were within the errors of those obtained via a 'gold standard' manual pathologist scoring.

Coulombic interactions between partially charged main-chain atoms not hydrogen-bonded to each other influence the conformations of alpha-helices and antiparallel beta-sheet. A new method for analysing the forces between hydrogen bonding groups in proteins includes all the Coulombic interactions.

An angle named gamma has been employed to describe the geometry at a hydrogen bond between main-chain atoms of polypeptides. In antiparallel beta-sheet, gamma is normally positive, whereas, in parallel beta-sheet and alpha-helices, it is negative. Although intriguing, no particular explanation has been offered to explain this result. We provide evidence that, in each case, the angular preference maximises the favourable Coulombic interaction between the partial negative charge on the carbonyl oxygen atom and the partial positive charge on the carbonyl carbon atom adjacent to the NH group to which it is hydrogen-bonded. Analyses of helices and beta-sheets in native proteins using Lennard-Jones potentials suggest that these carbonyl-carbonyl interactions are significant components of the attractive forces holding main-chain CONH groups together and are even in some cases larger than the hydrogen bonds themselves. A novel technique for analysing the forces holding together hydrogen-bonding groups in proteins is presented. It can be regarded as a development of the Kabsch and Sander method of calculating the energy of hydrogen bonds between main-chain atoms. In their program, electrostatic interactions are calculated between appropriate pairs of atoms, i.e. NH binding to CO. Instead, in our method, the four N, H, C, and O atoms, in a peptide bond are taken as a unit and the interaction between two NHCO groups calculated. We also use a Lennard-Jones potential, rather than just measuring the Coulombic interaction. With this approach, account is taken of all types of interactions between partially charged atoms, not only the hydrogen bonds.

Coulombic attractions between partially charged main-chain atoms stabilise the right-handed twist found in most beta-strands.

The use of Lennard-Jones potentials gives rise to an expected energy distribution for main-chain polypeptide conformations in the Ramachandran plot that matches well the observed distribution of phi, psi values in high-resolution proteins. The position of the energy minimum in the beta-strand conformation region is situated where there is a substantial contribution from the electrostatic attraction between the partial charge of the carbonyl carbon atom of one amino acid residue and that of the carbonyl oxygen atom of an adjacent residue. This attraction gives rise to a preference for the right-twisted beta-strand conformation compared with the left-twisted conformation. The majority of beta-sheets are twisted, almost always in one direction. Looking along a single strand, the twist is to the right. This twist also helps provide a rationale for the characteristic topology of the strand-helix-strand unit often observed in alpha/beta proteins. The electrostatic explanation for the twist we propose has not, to our knowledge, been explicitly suggested previously. The factor that has been most widely proposed to explain the twist is steric hindrance involving side-chain atoms. We provide evidence that the electrostatic effect is of comparable significance. Right-twisted beta-strands are geometrically closely related to polyproline II helices and to collagen helices, both of which are left-handed. Short regions of polyproline II type helices, which are sometimes, but not always, rich in proline residues, are common at protein surfaces. We point out that these helices are stabilised by the same carbonyl-carbonyl interactions as in right-twisted beta-strands.

Common ring motifs in proteins involving asparagine or glutamine amide groups hydrogen-bonded to main-chain atoms.

We report the frequent occurrence in proteins of motifs consisting of either 9-membered or 11-membered rings that involve the side-chain amide groups of asparagine and glutamine residues. The syn CO and NH groups of these amide groups are hydrogen-bonded to the main-chain NH and CO groups of other amino acid residues. The main-chain part of both the 9-membered and 11-membered rings has the conformation of a beta-strand. One such ring motifs occurs, on average, in half of all the proteins we examined. Similar conformations are found for most examples of the 9-membered and 11-membered rings. One of the 11-membered rings is distinct, compared to the others, in that its main-chain part has a mirror-image conformation. Another of the 11-membered rings occurs at the interior of the variable domains of some antibodies and assists in linking the two beta-sheets. We observe one 9-membered ring structure in a dihydrofolate reductase complex in which the amide in the nicotinamide group of the ligand NADP is bound to the enzyme. Groups that can form hydrogen bonds in a similar way to amide groups occur in several nucleotide bases; we find one example of a 9-membered ring involving adenine and main-chain atoms in the FAD-protein complex of glutathione reductase. Both have conformations like those of the other 9-membered rings.

Pyrrolidine ring puckering in cis and trans-proline residues in proteins and polypeptides. Different puckers are favoured in certain situations.

In a set of proteins studied at high resolution by X-ray crystallography over a half of all cis and trans-proline residues could be unambiguously assigned to one of the two forms of pyrrolidine ring puckering, called UP and DOWN. Of these, 89% of the cis-proline residues exhibit the DOWN pucker, while the trans-proline residues, on average, are about evenly distributed between the two forms. Of trans-proline residues found in alpha-helices, 79% have the UP ring pucker. trans-proline residues occurring in other situations are more equally distributed between the two forms of pucker, although further generalizations may be possible. Proline residues in a set of crystal structures of short polypeptides were also examined. As in the protein sample, a tendency for the cis-proline residues to have the DOWN pucker was observed, but the effect was less pronounced.