Biomedical Event Extraction with Machine Learning

Biomedical natural language processing (BioNLP) is a subfield of natural language processing, an area of computational linguistics concerned with developing programs that work with natural language: written texts and speech. Biomedical relation extraction concerns the detection of semantic relations such as protein-protein interactions (PPI) from scientific texts. The aim is to enhance information retrieval by detecting relations between concepts, not just individual concepts as with a keyword search. In recent years, events have been proposed as a more detailed alternative for simple pairwise PPI relations. Events provide a systematic, structural representation for annotating the content of natural language texts. Events are characterized by annotated trigger words, directed and typed arguments and the ability to nest other events. For example, the sentence “Protein A causes protein B to bind protein C” can be annotated with the nested event structure CAUSE(A, BIND(B, C)). Converted to such formal representations, the information of natural language texts can be used by computational applications. Biomedical event annotations were introduced by the BioInfer and GENIA corpora, and event extraction was popularized by the BioNLP'09 Shared Task on Event Extraction. In this thesis we present a method for automated event extraction, implemented as the Turku Event Extraction System (TEES). A unified graph format is defined for representing event annotations and the problem of extracting complex event structures is decomposed into a number of independent classification tasks. These classification tasks are solved using SVM and RLS classifiers, utilizing rich feature representations built from full dependency parsing. Building on earlier work on pairwise relation extraction and using a generalized graph representation, the resulting TEES system is capable of detecting binary relations as well as complex event structures. We show that this event extraction system has good performance, reaching the first place in the BioNLP'09 Shared Task on Event Extraction. Subsequently, TEES has achieved several first ranks in the BioNLP'11 and BioNLP'13 Shared Tasks, as well as shown competitive performance in the binary relation Drug-Drug Interaction Extraction 2011 and 2013 shared tasks. The Turku Event Extraction System is published as a freely available open-source project, documenting the research in detail as well as making the method available for practical applications. In particular, in this thesis we describe the application of the event extraction method to PubMed-scale text mining, showing how the developed approach not only shows good performance, but is generalizable and applicable to large-scale real-world text mining projects. Finally, we discuss related literature, summarize the contributions of the work and present some thoughts on future directions for biomedical event extraction. This thesis includes and builds on six original research publications. The first of these introduces the analysis of dependency parses that leads to development of TEES. The entries in the three BioNLP Shared Tasks, as well as in the DDIExtraction 2011 task are covered in four publications, and the sixth one demonstrates the application of the system to PubMed-scale text mining.

Variação sazonal em antioxidantes em folhas de plantas jovens de Caesalpinia echinata Lam. (pau-brasil)

O ambiente urbano pode gerar condições de estresse oxidativo nas plantas, levando à estimulação de antioxidantes como ácido ascórbico (AA), tióis (Ti), peroxidases (POD) e superóxido dismutase (SOD), em diferentes níveis, de acordo com sua capacidade de tolerar tais condições. Este estudo, em sua primeira fase, buscou conhecer o perfil destas defesas em folhas de plantas jovens de C. echinata com diferentes graus de desenvolvimento. Os níveis dos antioxidantes foram similares em todas as folhas analisadas. Em uma segunda fase, tais antioxidantes e a glutationa (GSH) foram avaliados em 130 plantas mantidas em casa de vegetação com monitoramento constante de temperatura, umidade relativa e irradiância, por um período de 18 meses, para estabelecer se variações sazonais nessas defesas ocorrem em resposta a fatores climáticos. Nesse período, os antioxidantes foliares foram analisados a cada três meses. Em geral, a concentração de AA e a atividade da POD foram altas em C. echinata, quando comparadas com as de outras espécies arbóreas tropicais. Houve variações nas defesas ao longo do tempo. A concentração de GSH e a atividade da POD mostraram ser influenciadas por mudanças na temperatura e a SOD foi estimulada em resposta à temperatura e à umidade relativa.

Absence of linkage between MHC and a gene involved in susceptibility to human schistosomiasis

Six hundred million people are at risk of infection by Schistosoma mansoni. MHC haplotypes have been reported to segregate with susceptibility to schistosomiasis in murine models. In humans, a major gene related to susceptibility/resistance to infection by S. mansoni (SM1) and displaying the mean fecal egg count as phenotype was detected by segregation analysis. This gene displayed a codominant mode of inheritance with an estimated frequency of 0.20-0.25 for the deleterious allele and accounted for more than 50% of the variance of infection levels. To determine if the SM1 gene segregates with the human MHC chromosomal region, we performed a linkage study by the lod score method. We typed for HLA-A, B, C, DR and DQ antigens in 11 informative families from an endemic area for schistosomiasis in Bahia, Brazil, by the microlymphocytotoxicity technique. HLA-DR typing by the polymerase chain reaction with sequence-specific primers (PCR-SSP) and HLA-DQ were confirmed by PCR-sequence-specific oligonucleotide probes (PCR-SSOP). The lod scores for the different q values obtained clearly indicate that there is no physical linkage between HLA and SM1 genes. Thus, susceptibility or resistance to schistosomiasis, as defined by mean fecal egg count, is not primarily dependent on the host's HLA profile. However, if the HLA molecule plays an important role in specific immune responses to S. mansoni, this may involve the development of the different clinical aspects of the disease such as granuloma formation and development of hepatosplenomegaly.

Cardiorespiratory adaptations induced by aerobic training in middle-aged men: the importance of a decrease in sympathetic stimulation for the contribution of dynamic exercise tachycardia

We investigated the effects of aerobic training on the efferent autonomic control of heart rate (HR) during dynamic exercise in middle-aged men, eight of whom underwent exercise training (T) while the other seven continued their sedentary (S) life style. The training was conducted over 10 months (three 1-h sessions/week on a field track at 70-85% of the peak HR). The contribution of sympathetic and parasympathetic exercise tachycardia was determined in terms of differences in the time constant effects on the HR response obtained using a discontinuous protocol (4-min tests at 25, 50, 100 and 125 watts on a cycle ergometer), and a continuous protocol (25 watts/min until exhaustion) allowed the quantification of the parameters (anaerobic threshold, VO2 AT; peak O2 uptake, VO2 peak; power peak) that reflect oxygen transport. The results obtained for the S and the T groups were: 1) a smaller resting HR in T (66 beats/min) when compared to S (84 beats/min); 2) during exercise, a small increase in the fast tachycardia (D0-10 s) related to vagal withdrawal (P<0.05, only at 25 watts) was observed in T at all powers; at middle and higher powers a significant decrease (P<0.05 at 50, 100 and 125 watts) in the slow tachycardia (D1-4 min) related to a sympathetic-dependent mechanism was observed in T; 3) the VO2 AT (S = 1.06 and T = 1.33 l/min) and VO2 peak (S = 1.97 and T = 2.47 l/min) were higher in T (P<0.05). These results demonstrate that aerobic training can induce significant physiological adaptations in middle-aged men, mainly expressed as a decrease in the sympathetic effects on heart rate associated with an increase in oxygen transport during dynamic exercise.

High dietary calcium decreases blood pressure in normotensive rats

This study evaluates the influence of different concentrations of calcium on blood pressure of normotensive rats. Four groups of Wistar rats (A, B, C and D) had free access to modified isocaloric and isoproteic diets containing 0.2, 0.5, 2 and 4 g% calcium as calcium carbonate for a period of 30 days. Systolic and diastolic arterial blood pressures were monitored in awake rats by the indirect tail cuff method using a Physiograph equipped with transducers and preamplifiers. Body weight and length and food intake were monitored. Under the conditions of the present experiment, the systolic and diastolic arterial blood pressures of group D rats fed a diet containing 4 g% calcium were significantly (P<0.05) lower compared to rats of the other groups.

Effect of all-trans retinoic acid on newly diagnosed acute promyelocytic leukemia patients: results of a Brazilian center

Thirty-seven patients with acute promyelocytic leukemia (APL) were treated with all-trans retinoic acid (ATRA). Patients received 45 mg m-2 day-1 po of ATRA until complete remission (CR) was achieved, defined as: a) presence of less than 5% blasts in the bone marrow, with b) white blood cells >103/mm3, c) platelets >105/mm3 and d) hemoglobin concentration >8 g/dl, with no blood or platelet transfusions. Thirty-one (83.7%) patients achieved CR by day 50, and 75% of these before day 30. Correction of the coagulopathy, achieved between days 2 and 10 (mean, 3 days), was the first evidence of response to treatment. Only one patient had been previously treated with chemotherapy and three had the microgranular variant M3 form. Dryness of skin and mucosae was the most common side effect observed in 82% of the patients. Thrombosis, hepatotoxicity and retinoid acid syndrome (RAS) were observed in 7 (19%), 6 (16%) and 4 (11%) patients, respectively. Thirteen (35%) patients had to be submitted to chemotherapy due to hyperleukocytosis (above 40 x 103/mm3) and six of these presented with new signs of coagulopathy after chemotherapy. Four (11%) patients died secondarily to intracerebral hemorrhage (IH) and two (5.4%) dropped out of the protocol due to severe ATRA side effects (one RAS and one hepatotoxicity). RAS and IH were related strictly to hyperleukocytosis. The reduced use of platelets and fresh frozen plasma probably lowered the total cost of treatment. We conclude that ATRA is an effective agent for inducing complete remission in APL patients.

Involvement of hippocampal NMDA receptors in retention of shuttle avoidance conditioning in rats

The purpose of this research was to evaluate the role of hippocampal N-methyl-D-aspartate (NMDA) receptors in acquisition and consolidation of memory during shuttle avoidance conditioning in rats. Adult male Wistar rats were surgically implanted with cannulae aimed at the CA1 area of the dorsal hippocampus. After recovery from surgery, animals were trained and tested in a shuttle avoidance apparatus (30 trials, 0.5-mA footshock, 24-h training-test interval). Immediately before or immediately after training, animals received a bilateral intrahippocampal 0.5-µl infusion containing 5.0 µg of the NMDA competitive receptor antagonist aminophosphonopentanoic acid (AP5) or vehicle (phosphate-buffered saline, pH 7.4). Infusion duration was 2 min per side. Pre-training infusion of AP5 impaired retention test performance (mean ± SEM number of conditioned responses (CRs) during retention test session was 16.47 ± 1.78 in the vehicle group and 9.93 ± 1.59 in the AP5 group; P<0.05). Post-training infusion of AP5 did not affect retention (mean ± SEM number of conditioned responses during retention test session was 18.46 ± 1.94 in the vehicle group and 20.42 ± 2.38 in the AP5 group; P>0.10). This impairment could not be attributed to an effect on acquisition, motor activity or footshock sensitivity since AP5 affected neither training session performance measured by the number of CRs nor the number of intertrial crossings during the training session. These data suggest that NMDA receptors in the hippocampus are critical for retention of shuttle avoidance conditioning, in agreement with previous evidence showing a role of NMDA receptors in fear memory.

Increased training prevents the impairing effect of intra-amygdala infusion of the non-NMDA receptor antagonist CNQX on inhibitory avoidance expression

Intra-amygdala infusion of the non-N-methyl-D-aspartate (NMDA) receptor antagonist 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX) prior to testing impairs inhibitory avoidance retention test performance. Increased training attenuates the impairing effects of amygdala lesions and intra-amygdala infusions of CNQX. The objective of the present study was to determine the effects of additional training on the impairing effects of intra-amygdala CNQX on expression of the inhibitory avoidance task. Adult female Wistar rats bilaterally implanted with cannulae into the border between the central and the basolateral nuclei of the amygdala were submitted to a single session or to three training sessions (0.2 mA, 24-h interval between sessions) in a step-down inhibitory avoidance task. A retention test session was held 48 h after the last training. Ten minutes prior to the retention test session, the animals received a 0.5-µl infusion of CNQX (0.5 µg) or its vehicle (25% dimethylsulfoxide in saline). The CNQX infusion impaired, but did not block, retention test performance in animals submitted to a single training session. Additional training prevented the impairing effect of CNQX. The results suggest that amygdaloid non-NMDA receptors may not be critical for memory expression in animals given increased training.

Disseminated Langerhans' cell histiocytosis and massive protein-losing enteropathy

Symptomatic involvement of the gastrointestinal (GI) tract as a prominent symptom in Langerhans' cell histiocytosis (LCH) is uncommon, occurring in less than 1 to 5% of all cases, even when the disease is in its disseminated form. Up to now, there have been reports of 18 cases of LCH with GI manifestations, including our 2 cases, with diarrhea (77.7%), protein-losing enteropathy (33.3%) and bloody stool being the most frequent findings. The authors present two patients with severe diarrhea and refractory hypoalbuminemia, and with the protein-losing enteropathy documented by Cr51-labeled albumin studies. A review of the literature indicated that the presence of GI symptoms is often associated with systemic disease as well as with poor prognosis, mainly under 2 years of age. Radioisotopes are useful for documenting protein loss in several diseases with high specificity and sensitivity, and their utilization in the cases reviewed here permitted diagnoses in 6 children, as well as improved therapeutic management.

Prevalence and risk factors for HBV, HCV and HDV infections among injecting drug users from Rio de Janeiro, Brazil

Viral hepatitis constitutes a major health issue, with high prevalence among injecting drug users (IDUs). The present study assessed the prevalence and risk determinants for hepatitis B, C and D viruses (HBV, HCV and HDV) infections among 102 IDUs from Rio de Janeiro, Brazil. Serological markers and HCV-RNA were detected by enzyme immunoassay and nested PCR, respectively. HCV genotyping was determined by restriction fragment length polymorphism analysis (RFLP). HBsAg, anti-HBc and anti-HBs were found in 7.8, 55.8 and 24.7% of IDUs, respectively. In the final logistic regression, HBV infection was independently associated with male homosexual intercourse within the last 5 years (odds ratio (OR) 3.1; 95% confidence interval (CI) 1.1-8.8). No subject presented anti-delta (anti-HD). Anti-HCV was detected in 69.6% of subjects, and was found to be independently associated with needle sharing in the last 6 months (OR 3.4; 95% CI 1.3-9.2) and with longer duration of iv drug use (OR 3.1; 95% CI 1.1-8.7). These data demonstrate that this population is at high risk for both HBV and HCV infection. Among IDUs from Rio de Janeiro, unprotected sexual intercourse seems to be more closely associated with HBV infection, whereas HCV is positively correlated with high risk injecting behavior. Comprehensive public health interventions targeting this population and their sexual partners must be encouraged.

A pulsatile flow model for in vitro quantitative evaluation of prosthetic valve regurgitation

A pulsatile pressure-flow model was developed for in vitro quantitative color Doppler flow mapping studies of valvular regurgitation. The flow through the system was generated by a piston which was driven by stepper motors controlled by a computer. The piston was connected to acrylic chambers designed to simulate "ventricular" and "atrial" heart chambers. Inside the "ventricular" chamber, a prosthetic heart valve was placed at the inflow connection with the "atrial" chamber while another prosthetic valve was positioned at the outflow connection with flexible tubes, elastic balloons and a reservoir arranged to mimic the peripheral circulation. The flow model was filled with a 0.25% corn starch/water suspension to improve Doppler imaging. A continuous flow pump transferred the liquid from the peripheral reservoir to another one connected to the "atrial" chamber. The dimensions of the flow model were designed to permit adequate imaging by Doppler echocardiography. Acoustic windows allowed placement of transducers distal and perpendicular to the valves, so that the ultrasound beam could be positioned parallel to the valvular flow. Strain-gauge and electromagnetic transducers were used for measurements of pressure and flow in different segments of the system. The flow model was also designed to fit different sizes and types of prosthetic valves. This pulsatile flow model was able to generate pressure and flow in the physiological human range, with independent adjustment of pulse duration and rate as well as of stroke volume. This model mimics flow profiles observed in patients with regurgitant prosthetic valves.

Angiotensin-(1-7) potentiates the coronary vasodilatatory effect of bradykinin in the isolated rat heart

It has been shown that angiotensin-(1-7) (Ang-(1-7)) infusion potentiates the bradykinin (BK)-induced hypotensive response in conscious rats. The present study was conducted to identify Ang-(1-7)-BK interactions in the isolated rat heart perfused according to the Langendorff technique. Hearts were excised and perfused through the aortic stump under a constant flow with Krebs-Ringer solution and the changes in perfusion pressure and heart contractile force were recorded. Bolus injections of BK (2.5, 5, 10 and 20 ng) produced a dose-dependent hypotensive effect. Ang-(1-7) added to the perfusion solution (2 ng/ml) did not change the perfusion pressure or the contractile force but doubled the hypotensive effect of the lower doses of BK. The BK-potentiating Ang-(1-7) activity was blocked by pretreatment with indomethacin (5 mg/kg, ip) or L-NAME (30 mg/kg, ip). The Ang-(1-7) antagonist A-779 (50 ng/ml in Krebs-Ringer) completely blocked the effect of Ang-(1-7) on BK-induced vasodilation. These data suggest that the potentiation of the BK-induced vasodilation by Ang-(1-7) can be attributed to the release of nitric oxide and vasodilator prostaglandins through an Ang-(1-7) receptor-mediated mechanism.

Structural, functional and immunological studies on a polymeric bacterial protein

The characterization of proteins from Brucella spp, the causative agent of brucellosis, has been the subject of intensive research. We have described an 18-kDa cytoplasmic protein of Brucella abortus and shown the potential usefulness of this protein as an antigen for the serologic diagnosis of brucellosis. The amino acid sequence of the protein showed a low but significant homology with that of lumazine synthases. Lumazine is an intermediate product in bacterial riboflavin biosynthesis. The recombinant form of the 18-kDa protein (expressed in E. coli) folds like the native Brucella protein and has lumazine-synthase enzymatic activity. Three-dimensional analysis by X-ray crystallography of the homolog Bacillus subtilis lumazine synthase has revealed that the enzyme forms an icosahedral capsid. Recombinant lumazine synthase from B. abortus was crystallized, diffracted X rays to 2.7-Å resolution at room temperature, and the structure successfully solved by molecular replacement procedures. The macromolecular assembly of the enzyme differs from that of the enzyme from B. subtilis. The Brucella enzyme remains pentameric (90 kDa) in its crystallographic form. Nonetheless, the active sites of the two enzymes are virtually identical at the structural level, indicating that inhibitors of these enzymes could be viable pharmaceuticals across a broad species range. We describe the structural reasons for the differences in their quaternary arrangement and also discuss the potential use of this protein as a target for the development of acellular vaccines.

Amplification of 9q34 in childhood adrenocortical tumors: a specific feature unrelated to ethnic origin or living conditions

Adrenocortical tumors (ACT) in children under 15 years of age exhibit some clinical and biological features distinct from ACT in adults. Cell proliferation, hypertrophy and cell death in adrenal cortex during the last months of gestation and the immediate postnatal period seem to be critical for the origin of ACT in children. Studies with large numbers of patients with childhood ACT have indicated a median age at diagnosis of about 4 years. In our institution, the median age was 3 years and 5 months, while the median age for first signs and symptoms was 2 years and 5 months (N = 72). Using the comparative genomic hybridization technique, we have reported a high frequency of 9q34 amplification in adenomas and carcinomas. This finding has been confirmed more recently by investigators in England. The lower socioeconomic status, the distinctive ethnic groups and all the regional differences in Southern Brazil in relation to patients in England indicate that these differences are not important to determine 9q34 amplification. Candidate amplified genes mapped to this locus are currently being investigated and Southern blot results obtained so far have discarded amplification of the abl oncogene. Amplification of 9q34 has not been found to be related to tumor size, staging, or malignant histopathological features, nor does it seem to be responsible for the higher incidence of ACT observed in Southern Brazil, but could be related to an ACT from embryonic origin.

Adrenocortical tumors in children

Childhood adrenocortical tumors (ACT) are rare. In the USA, only about 25 new cases occur each year. In Southern Brazil, however, approximately 10 times that many cases are diagnosed each year. Most cases occur in the contiguous states of São Paulo and Paraná. The cause of this higher rate has not been identified. Familial genetic predisposition to cancer (p53 mutations) and selected genetic syndromes (Beckwith-Wiedemann syndrome) have been associated with childhood ACT in general but not with the Brazilian counterpart. Most of the affected children are young girls with classic endocrine syndromes (virilizing and/or Cushing). Levels of urinary 17-ketosteroids and plasma dehydroepiandrosterone sulfate (DHEA-S), which are abnormal in approximately 90% of the cases, provide the pivotal clue to a diagnosis of ACT. Typical imaging findings of pediatric ACT consist of a large, well-defined suprarenal tumor containing calcifications with a thin capsule and central necrosis or hemorrhage. The pathologic classification of pediatric ACT is troublesome. Even an experienced pathologist can find it difficult to differentiate carcinoma from adenoma. Surgery is the single most important procedure in the successful treatment of ACT. The role of chemotherapy in the management of childhood ACT has not been established although occasional tumors are responsive to mitotane or cisplatin-containing regimens. Because of the heterogeneity and rarity of the disease, prognostic factors have been difficult to establish in pediatric ACT. Patients with incomplete tumor resection or with metastatic disease at diagnosis have a dismal prognosis. In patients with localized and completely resected tumors, the size of the tumor has predictive value. Patients with large tumors have a much higher relapse rate than those with small tumors.