PreBIND and Textomy – mining the biomedical literature for protein-protein interactions using a support vector machine

Background

The majority of experimentally verified molecular interaction and biological pathway data are present in the unstructured text of biomedical journal articles where they are inaccessible to computational methods. The Biomolecular interaction network database (BIND) seeks to capture these data in a machine-readable format. We hypothesized that the formidable task-size of backfilling the database could be reduced by using Support Vector Machine technology to first locate interaction information in the literature. We present an information extraction system that was designed to locate protein-protein interaction data in the literature and present these data to curators and the public for review and entry into BIND.

Results

Cross-validation estimated the support vector machine's test-set precision, accuracy and recall for classifying abstracts describing interaction information was 92%, 90% and 92% respectively. We estimated that the system would be able to recall up to 60% of all non-high throughput interactions present in another yeast-protein interaction database. Finally, this system was applied to a real-world curation problem and its use was found to reduce the task duration by 70% thus saving 176 days.

Conclusions

Machine learning methods are useful as tools to direct interaction and pathway database back-filling; however, this potential can only be realized if these techniques are coupled with human review and entry into a factual database such as BIND. The PreBIND system described here is available to the public at http://bind.ca. Current capabilities allow searching for human, mouse and yeast protein-interaction information.     

Cross-adaptation and molecular modeling study of receptor mechanisms common to four taste stimuli in humans

Psychophysical cross-adaptation experiments were performed with two carbohydrates, sucrose (SUC) and fructose (FRU), and two sweeteners, acesulfame-K (MOD) and dulcin (DUL). Seven subjects were asked to match concentrations that elicited the same intensity as a sucrose reference (30 g/l). Cross-adaptation levels were calculated as the ratio of isointense concentrations measured for a given stimulus before and under adaptation. On average, cross-adaptation between SUC and FRU is low and apparently reciprocal. By contrast, cross-adaptation between SUC and MOD is clearly non-reciprocal: SUC adapts MOD significantly (24%, P < 0.005), but MOD fails to adapt SUC (2%, P < 0.79). Significant and reciprocal cross-enhancement is observed between DUL and MOD (approximately -20%, P < 0.03), and also between SUC and DUL (approximately -15%, P < 0.08). In parallel, molecular modeling of the four tastants was performed in order to look for the 12 common binding motifs that were isolated on 14 other tastants in a previous study. SUC and FRU each display 10 out of the 12 binding motifs, whereas DUL and MOD only display four and five distinct motifs respectively and do not have any motif in common. Experimental cross-adaptation levels seem to correlate well with the number of motifs that molecules have in common. FRU and SUC share a majority of binding motifs and correlatively show mutual cross-adaptation. Four motifs of MOD are found among the 10 motifs of SUC, which may explain why SUC cross-adapts MOD but not vice versa. By contrast, DUL and MOD do not share any motif and do not cross- adapt. The various molecular mechanisms that may be responsible for cross-adaptation and/or cross-enhancement are discussed in light of our results.      

Muscle glycogen depletion and subsequent replenishment affect anaerobic capacity of horses.

The purpose of this study was to determine the effect of muscle glycogen depletion and subsequent replenishment on anaerobic capacity of horses. In a blinded crossover study, seven fit horses performed glycogen-depleting exercise on two occasions. Horses were infused after glycogen-depleting exercise with either 6 g/kg body wt of glucose as a 13.5% solution in 0.9% NaCl (Glu) or with 0.9% NaCl (Sal) of equivalent volume. Subsequently, horses performed a high-speed exercise test (120% of maximal rate of oxygen consumption) to estimate maximum accumulated oxygen deficit. Replenishment of muscle glycogen was greater (P < 0.05) in Glu [from 24.7 +/- 7.2 (SE) to 116.5 +/- 7 mmol/kg wet wt before and after infusion, respectively] than in Sal (from 23.4 +/- 7.2 to 47.8 +/- 5.7 mmol/kg wet wt before and after infusion, respectively). Run time to fatigue during the high-speed exercise test (97.3 +/- 8.2 and 70.8 +/- 8.3 s, P < 0.05), maximal accumulated oxygen deficit (105.7 +/- 9.3 and 82.4 +/- 10.3 ml O(2) equivalent/kg, P < 0.05), and blood lactate concentration at the end of the high-speed exercise test (11.1 +/- 1.4 and 9.2 +/- 3.7 mmol/l, P < 0.05) were greater for Glu than for Sal, respectively. We concluded that decreased availability of skeletal muscle glycogen stores diminishes anaerobic power generation and capacity for high-intensity exercise in horses.     

Immunocytochemistry with osmium-fixed tissue. I. Light microscopic localization of growth hormone and prolactin with the unlabeled antibody-enzyme method.

Growth hormone and prolactin were localized on thin plastic sections of rat anterior pituitary gland and mammosomatotropic tumor MtTW15 that were fixed with osmium tetroxide (alone,mixed with aldehydes, or after aldehydes). Intense immunocytochemical staining for both antigens was obtained after plastic was removed from sections with an alcoholic solution of sodium hydroxide. The results indicated that antigenic determinants of rat prolactin and growth hormone were not completely destroyed or inactivated by fixation with osmium and embedment in epoxy resin, and that removal of the polymerized epoxy resin was necessary to obtain light microscopic postembedding immunocytochemical staining of these antigens. The results also demonstrated that tissues which have been conventionally processed for morphological evaluation by electron microscopy may be suitable for postembedding immunocytochemical staining of some antigens for light microscopy.     

DIFFERENTIATION OF NEUROBLASTOMA, GLIOMA, AND HYBRID CELLS IN CULTURE AS MEASURED BY THE SYNTHESIS OF SPECIFIC PROTEIN SPECIES: EVIDENCE FOR NEUROBLAST-GLIOBLAST RECIPROCAL GENETIC REGULATION

Abstract— Protein species from differentiating neuroblastoma, glioma, and hybrid neuroblastoma-glioma cell lines in cell culture were separated and identified initially in the first dimension by the use of isoelectric focusing gels and were further separated in the second dimension by SDS-acrylamide gels. There were two main classes of proteins identified: proteins which were dominantly expressed in neuroblastoma and also in hybrid cell cultures, and proteins which were expressed in glioma and also hybrid cell cultures. In general, proteins were identified which were significantly expressed in neuroblastoma cells and much reduced in glioma cultures, and also conversely so. The hybrid cell line expressed many of the neuroblastoma-type proteins and relatively fewer of the glioma type proteins. A specific protein species (2) was identified in hybrid cells and was not present in either parental neuroblastoma or glioma cultures. Protein z was expressed however by the co-culturing of neuroblastoma and glioma cells suggesting its induction is dependent on a soluble factor. Protein z in hybrid cells was demonstrated in both stained gels and by autoradiography. Chromosome analysis of hybrid cells confirmed the presence of both rat and mouse chromosomes. It is suggested that similar neuronal-glial interaction may be functional in the intact brain, and that similar reciprocal modulation between neurons and glia may be a central mechanism of differentiation in the nervous system.     

DORMANCY AND GERMINATION IN SEEDS OF COMMON RAGWEED WITH REFERENCE TO BEAL'S BURIED SEED EXPERIMENT

Common ragweed (Ambrosia artemisiifolia L.) was one of 19 herbaceous weedy species used by Beal in his buried viable seed experiment started in 1879. No seeds germinated during the first 35 years of the experiment when germination tests were performed in late spring, summer or early autumn. Germination did occur in seeds buried for 40 years when seeds were exhumed and tested for germination in early spring. Data obtained in more recent research provide the probable explanation for these results. Seeds of common ragweed that do not germinate in spring enter secondary dormancy by mid to late spring and will not germinate until dormancy is broken the following late autumn and winter. Thus, during the first 35 years of the experiment seeds were dormant when tested for germination, whereas seeds buried for 40 years were nondormant. Seeds buried 50 years or longer did not germinate when tested in spring, probably because they had lost viability and/or seeds germinated during burial and seedlings died.     

Mitochondrial growth and division during the cell cycle in HeLa cells

The growth and division of mitochondria during the cell cycle was investigated by a morphometric analysis of electron micrographs of synchronized HeLa cells. The ratio of total outer membrane contour length to cytoplasmic area did not vary significantly during the cell cycle, implying a continuous growth of the mitochondrial outer membrane. The mean fraction of cytoplasmic area occupied by mitochondrial profiles was likewise found to remain constant, indicating that the increase in total mitochondrial volume per cell occurs continuously during interphase, in such a way that the mitochondrial complement occupies a constant fraction( approximately 10-11(percent)) of the volume of the cytoplasm. The mean area, outer membrane contour length, and axis ratio of the mitochondrial profiles also did not vary appreciably during the cell cycle; furthermore, the close similarity of the frequency distributions of these parameters for the six experimental time-points suggested a stable mitochondrial shape distribution. The constancy of both the mean mitochondrial profile area and the number of mitochondrial profiles per unit of cytoplasmic area was interpreted to indicate the continuous division of mitochondria at the level of the cell population. Furthermore, no evidence was found for the occurrence of synchronous mitochondrial growth and division within individual cells. Thus, it appears that, in HeLa cells, there is no fixed temporal relationship between the growth and division of mitochondria and the events of the cell cycle. A number of statistical methods were developed for the purpose of making numerical estimates of certain three-dimensional cellular and mitochondrial parameters. Mean cellular and cytoplasmic volumes were calculated for the six time-points; both exhibited a nonlinear, approx. twofold increase. A comparison of the axis ratio distributions of the mitochondrial profiles with theoretical distributions expected from random sectioning of bodies of various three-dimensional shapes allowed the derivation of an "average" mitochondrial shape. This, in turn, permitted calculations to be made which expressed the two-dimensional results in three-dimensional terms. Thus, the estimated values for the number of mitochondria per unit of cytoplasmic volume and for the mean mitochondrial volume were found to remain constant during the cell cycle, while the estimated number of mitochondria per cell increase approx. twofold in an essentially continuous manner.     

Comparison of the (Ca2+ + Mg2+)-ATPase proteins from normal and dystrophic chicken sarcoplasmic reticulum.

The involvement of membrane protein in dystrophic chicken fragmented sarcoplasmic reticulum alterations has been examined. A purified preparation of the (Ca2+ + Mg2+)-ATPase protein from dystrophic fragmented sarcoplasmic reticulum was found to have a reduced calcium-sensitive ATPase activity and phosphoenzyme level, in agreement with alterations found in dystrophic chicken fragmented sarcoplasmic reticulum. An amino acid analysis of the ATPase preparations showed no difference in the normal and dystrophic (Ca2+ + Mg2+)-ATPase. The (Ca2+ + Mg2+)-ATPase was investigated further by isoelectric focusing and proteolytic digestion of the fragmented sarcoplasmic reticulum. Neither of these methods indicated any alteration in the composition of the dystrophic (Ca2+ + Mg2+)-ATPase. We have concluded that the alterations observed in dystrophic fragmented sarcoplasmic reticulum are not due to increased amounts of non-(Ca2+ + Mg2+)-ATPase protein, and that the normal and dystrophic (Ca2+ + Mg2+)-ATPase protein are not detectably different.     

Microtubules and cyclic amp in human leukocytes: on the order of things

We have shown previously that the β-adrenergic agonist isoproterenol (2μM) and the phosphodiesterase inhibitor isobutylmethylxanthine (1 mM) produce a much greater increase in cyclic AMP in human leukocytes that have been pretreated with colchicine (or with other agents that affect microtubule assembly) than in control leukocytes. The effects of colchicines were both time- and dose-dependant. These and other data suggested that the generation of cyclic AMP is normally restricted by an intact system of cytoplasmic microtubules. If so, then the same time and dose dependencies might apply to other colchicines-induced changes in leukocyte function. We have now assayed the distribution of concanavalin A (Con A)-receptor complexes on the leukocyte membrane, taking into account that leukocytes competent to assemble microtubules show a uniform distribution of surface- bound Con A whereas microtubule-deficient cells accumulate Con A in surface caps. We have found that the effect of colchicine on capping is also both time- and dose dependent, and that the dose-response relationships conform to those required to increase cyclic AMP levels. These findings provide further evidence that both colchicine-induced Con-A capping and colchicine- induced cyclic AMP generation depend upon the relaxation of constraints normally imposed by cytoplasmic microtubules upon the plasma membrane, which limit, respectively, lateral mobility of the lectin-receptor complexes, and expression of hormone-sensitive adenylate cyclase. Moreover, colchicine-induced Con-A cap formation is not affected even by very large changes in leukocyte cyclic AMP levels. Thus, elevated cyclic AMP levels do not appear to promote the dissolution of microtubules; rather, the dissolution of microtubules permits the generation of increased amounts of cyclic AMP.     

A light-scattering measurement of membrane vesicle permeability.

Light-scattering/intensity autocorrelation measurements of vesicle diffusivity were used to follow the time course of the osmotic response of lobster abdominal sarcoplasmic reticulum vesicles to five lipophobic nonelectrolytes. Steady-state portions of the resulting time traces show these vesicles to be permeable to ethylene glycol and glycerol and impermeable to erythritol, glucose, and sucrose. Using measured values of the hydrodynamic radii of these nonelectrolytes, it is concluded that under passive transport conditions, these vesicles may be thought of as having pores whose radii lie between 3.1 and 3.5 A. In addition, the results presented here indicated that above a certain impermeable nonelectrolyte concentration, vesicles did not respond osmotically even though they had not collapsed. This suggests that at least under the experimental conditions reported here, vesicles behaved as if rigid when their average volume had decreased to about 50% of its original isotonic value.