Innovative experimental animal models for real-time comparison of antithrombogenicity between two oxygenators using dual extracorporeal circulation circuits and indocyanine green fluorescence imaging.

BACKGROUND: Antithrombogenicity of extracorporeal membrane oxygenation (ECMO) devices, particularly oxygenators, is a current problem, with numerous studies and developments underway. However, there has been limited progress in developing methods to accurately compare the antithrombogenicity of oxygenators. Animal experiments are commonly conducted to evaluate the antithrombogenicity of devices; however, it is challenging to maintain a steady experimental environment. We propose an innovative experimental animal model to evaluate different devices in a constant experimental environment in real-time. METHODS: This model uses two venous-arterial ECMO circuits attached to one animal (one by jugular vein and carotid artery, one by femoral vein and artery) and real-time assessment of thrombus formation in the oxygenator by indocyanine green (ICG) fluorescence imaging. Comparison studies were conducted using three pigs: one to compare different oxygenators (MERA vs. CAPIOX) (Case 1), and two to compare antithrombotic properties of the oxygenator (QUADROX) when used under different hydrodynamic conditions (continuous flow vs. pulsatile flow) (Cases 2 and 3). RESULTS: Thrombi, visualized using ICG imaging, appeared as black dots on a white background in each oxygenator. In Case 1, differences in the site of thrombus formation and rate of thrombus growth were observed in real-time in two oxygenators. In Case 2 and 3, the thrombus region was smaller in pulsatile than in continuous conditions. CONCLUSIONS: We devised an innovative experimental animal model for comparison of antithrombogenicity in ECMO circuits. This model enabled simultaneous evaluation of two different ECMO circuits under the same biological conditions and reduced the number of sacrificed experimental animals.

CMIC: predicting DNA methylation inheritance of CpG islands with embedding vectors of variable-length k-mers.

BACKGROUND: Epigenetic modifications established in mammalian gametes are largely reprogrammed during early development, however, are partly inherited by the embryo to support its development. In this study, we examine CpG island (CGI) sequences to predict whether a mouse blastocyst CGI inherits oocyte-derived DNA methylation from the maternal genome. Recurrent neural networks (RNNs), including that based on gated recurrent units (GRUs), have recently been employed for variable-length inputs in classification and regression analyses. One advantage of this strategy is the ability of RNNs to automatically learn latent features embedded in inputs by learning their model parameters. However, the available CGI dataset applied for the prediction of oocyte-derived DNA methylation inheritance are not large enough to train the neural networks. RESULTS: We propose a GRU-based model called CMIC (CGI Methylation Inheritance Classifier) to augment CGI sequence by converting it into variable-length k-mers, where the length k is randomly selected from the range [Formula: see text] to [Formula: see text], N times, which were then used as neural network input. N was set to 1000 in the default setting. In addition, we proposed a new embedding vector generator for k-mers called splitDNA2vec. The randomness of this procedure was higher than the previous work, dna2vec. CONCLUSIONS: We found that CMIC can predict the inheritance of oocyte-derived DNA methylation at CGIs in the maternal genome of blastocysts with a high F-measure (0.93). We also show that the F-measure can be improved by increasing the parameter N, that is, the number of sequences of variable-length k-mers derived from a single CGI sequence. This implies the effectiveness of augmenting input data by converting a DNA sequence to N sequences of variable-length k-mers. This approach can be applied to different DNA sequence classification and regression analyses, particularly those involving a small amount of data.

A convolutional neural network-based regression model to infer the epigenetic crosstalk responsible for CG methylation patterns.

BACKGROUND: Epigenetic modifications, including CG methylation (a major form of DNA methylation) and histone modifications, interact with each other to shape their genomic distribution patterns. However, the entire picture of the epigenetic crosstalk regulating the CG methylation pattern is unknown especially in cells that are available only in a limited number, such as mammalian oocytes. Most machine learning approaches developed so far aim at finding DNA sequences responsible for the CG methylation patterns and were not tailored for studying the epigenetic crosstalk. RESULTS: We built a machine learning model named epiNet to predict CG methylation patterns based on other epigenetic features, such as histone modifications, but not DNA sequence. Using epiNet, we identified biologically relevant epigenetic crosstalk between histone H3K36me3, H3K4me3, and CG methylation in mouse oocytes. This model also predicted the altered CG methylation pattern of mutant oocytes having perturbed histone modification, was applicable to cross-species prediction of the CG methylation pattern of human oocytes, and identified the epigenetic crosstalk potentially important in other cell types. CONCLUSIONS: Our findings provide insight into the epigenetic crosstalk regulating the CG methylation pattern in mammalian oocytes and other cells. The use of epiNet should help to design or complement biological experiments in epigenetics studies.

Suitable hemolysis index for low-flow rotary blood pumps.

A suitable index is needed for hemolysis tests that use low-flow pumps, such as pediatric blood pumps or blood purification pumps. To create such an index, the present study investigates the change of plasma-free hemoglobin in the pump circuit with time and the change of the hemolysis rate with flow rate and impeller rotational speed. The results show that the hemolysis rate or the increase rate of the total free hemoglobin are suitable measures for hemolysis evaluation for low-flow pumps.

Plasma skimming efficiency of human blood in the spiral groove bearing of a centrifugal blood pump.

This work investigates the plasma skimming effect in a spiral groove bearing within a hydrodynamically levitated centrifugal blood pump when working with human blood having a hematocrit value from 0 to 40%. The present study assessed the evaluation based on a method that clarified the limitations associated with such assessments. Human blood was circulated in a closed-loop circuit via a pump operating at 4000 rpm at a flow rate of 5 L/min. Red blood cells flowing through a ridge area of the bearing were directly observed using a high-speed microscope. The hematocrit value in the ridge area was calculated using the mean corpuscular volume, the bearing gap, the cross-sectional area of a red blood cell, and the occupancy of red blood cells. The latter value was obtained from photographic images by dividing the number of pixels showing red blood cells in the evaluation area by the total number of pixels in this area. The plasma skimming efficiency was calculated as the extent to which the hematocrit of the working blood was reduced in the ridge area. For the hematocrit in the circuit from 0 to 40%, the plasma skimming efficiency was approximately 90%, meaning that the hematocrit in the ridge area became 10% as compared to that in the circuit. For a hematocrit of 20% and over, red blood cells almost completely occupied the ridge. Thus, a valid assessment of plasma skimming was only possible when the hematocrit was less than 20%.

Improvement of hemolysis performance in a hydrodynamically levitated centrifugal blood pump by optimizing a shroud size.

We have developed a hydrodynamically levitated centrifugal blood pump. In the blood pump having hydrodynamic bearings, the narrow bearing gap has a potential for high hemolysis. The purpose of the this study is to improve hemolysis performance in a hydrodynamically levitated centrifugal blood pump by optimizing a shroud size. The impeller was levitated passively at the position where the thrust forces acting on the impeller were balanced. We focused on a size of a bottom shroud with a hydrodynamic bearing that could change the bottom hydrodynamic force to balance the thrust force at the wide bearing gap for reducing hemolysis. Five test models with various shroud size were compared: 989 mm2 (HH-10.5), 962 mm2 (HH-12), 932 mm2 (HH-13.5), 874 mm2 (HH-16), and 821 mm2 (HH-18). A numerical analysis was first performed to estimate the bearing gaps in the test model. The bearing gaps were then measured to validate the numerical analysis. Finally, an in vitro hemolysis test was performed. The numerical analysis revealed that the HH-13.5 model had the widest bearing gap of 129 µm. In the measurement test, the estimation error for the bearing gap was less than 10%. In the hemolysis test, the HH-13.5 model achieved the lowest hemolysis level among the five models. The present study demonstrated that the numerical analysis was found to be effective for determining the optimal should size, and the HH-13.5 model had the optimal shroud size in the developed hydrodynamically levitated centrifugal blood pump to reduce hemolysis.

Detection of thrombosis in a magnetically levitated blood pump by vibrational excitation of the impeller.

The use of contactless support technology for the impeller has led to an increase in the durability of ventricular assist devices (VADs), and these have been in clinical use worldwide. However, pump thrombosis and stroke are still issues to be solved. We have developed a method for detecting the thrombosis in a magnetically levitated blood pump without the need for additional sensors or other equipment. In the proposed method, a sinusoidal current is applied to the electromagnets used for the magnetic bearing, resulting in vibration of the impeller. The phase difference between the current and displacement of the impeller increases with pump thrombosis. First, we describe the principle by which the pump thrombosis is detected. Pump thrombosis reduces the narrowest fluid gap in the pump and this gives rise to a change in the phase difference. Second, we report on experiments in which we changed the narrowest fluid gap using oriented polypropylene tape and showed that decreasing the narrowest fluid gap resulted in an increase in phase difference. For these experiments, the measurements were repeated three times for each condition. Third, we examine the relationship between the pump thrombosis and the phase difference evaluated by observations of the underside of the impeller when operating the pump with porcine blood. Since light was unable to penetrate the blood layer, the erythrocytes were removed for this observation. Only one observation was made. The results showed the phase difference rapidly increased at the same moment when the pump thrombosis was observed. This implies the proposed method has the potential to detect the early stages of pump thrombosis. Finally, in vitro experiments to detect thrombosis when using whole porcine blood in the pump were conducted. The experiment was carried out five times. To intentionally form a thrombus inside the pump, the activated clotting time was controlled to be less than 200 s. In every case, the phase difference increased by more than one degree after tens of minutes. Then, the pump was disassembled and a small amount of pump thrombosis was observed. We conclude that real-time diagnosis of pump thrombosis may be realized by measuring the phase difference without the need for additional sensors.

RocSampler: regularizing overlapping protein complexes in protein-protein interaction networks.

BACKGROUND: In recent years, protein-protein interaction (PPI) networks have been well recognized as important resources to elucidate various biological processes and cellular mechanisms. In this paper, we address the problem of predicting protein complexes from a PPI network. This problem has two difficulties. One is related to small complexes, which contains two or three components. It is relatively difficult to identify them due to their simpler internal structure, but unfortunately complexes of such sizes are dominant in major protein complex databases, such as CYC2008. Another difficulty is how to model overlaps between predicted complexes, that is, how to evaluate different predicted complexes sharing common proteins because CYC2008 and other databases include such protein complexes. Thus, it is critical how to model overlaps between predicted complexes to identify them simultaneously. RESULTS: In this paper, we propose a sampling-based protein complex prediction method, RocSampler (Regularizing Overlapping Complexes), which exploits, as part of the whole scoring function, a regularization term for the overlaps of predicted complexes and that for the distribution of sizes of predicted complexes. We have implemented RocSampler in MATLAB and its executable file for Windows is available at the site, http://imi.kyushu-u.ac.jp/~om/software/RocSampler/ . CONCLUSIONS: We have applied RocSampler to five yeast PPI networks and shown that it is superior to other existing methods. This implies that the design of scoring functions including regularization terms is an effective approach for protein complex prediction.

Optical Dynamic Analysis of Thrombus Inside a Centrifugal Blood Pump During Extracorporeal Mechanical Circulatory Support in a Porcine Model.

Complications due to pump thrombus remain the weak point of mechanical circulatory support (MCS), such as the use of a left ventricular assist device (LVAD) or extracorporeal membrane oxygenation, leading to poor outcomes. Hyperspectral imaging (HSI) is an effective imaging method using a hyperspectral (HS) camera, which comprises a spectrophotometer and a charge-coupled device camera to discriminate thrombus from whole blood. Animal experiments were conducted to analyze dynamic imaging of thrombus inside a prototype of a hydrodynamically levitated centrifugal blood pump using an HSI system. Six pigs were divided into a venous circulation group (n = 3) and an arterial circulation group (n = 3). Inflow and outflow cannulae were inserted into the jugular veins in the venous circulation group. The latter simulated an LVAD application. To create thrombogenic conditions, pump flow was maintained at 1 L/min without anticoagulation. An image of the bottom surface of the pump was captured by the HS camera every 4 nm over the wavelength range of 608-752 nm. Real-time dynamic images of the inside of the pump were displayed on the monitor. Appearance of an area displaying thrombus was detected within 24 h after the start of the circulation in every experiment. This imaging system also succeeded in determining the origins of pump thrombus: from inside the pump in two cases, and from outside in four cases. Two main possible sources of pump thrombus originating outside the pump were identified on autopsy: wedge thrombus around the inflow cannula; and string-like thrombus at the junction between the pump inlet and circuit tube. The results of this study from close observation of the changing appearance of pump thrombus may contribute to improvements in the safety of extracorporeal MCS.

Long-term durability test of axial-flow ventricular assist device under pulsatile flow.

A long-term durability test was conducted on a newly developed axial-flow ventricular assist device (VAD) with hydrodynamic bearings. The mock circulatory loop consisted of a diaphragm pump with a mechanical heart valve, a reservoir, a compliance tank, a resistance valve, and flow paths made of polymer or titanium. The VAD was installed behind the diaphragm pump. The blood analog fluid was a saline solution with added glycerin at a temperature of 37 °C. A pulsatile flow was introduced into the VAD over a range of flow rates to realize a positive flow rate and a positive pressure head at a given impeller rotational speed, yielding a flow rate of 5 L/min and a pressure of 100 mmHg. Pulsatile flow conditions were achieved with the diastolic and systolic flow rates of ~0 and 9.5 L/min, respectively, and an average flow rate of ~5 L/min at a pulse rate of 72 bpm. The VAD operation was judged by not only the rotational speed of the impeller, but also the diastolic, systolic, and average flow rates and the average pressure head of the VAD. The conditions of the mock circulatory loop, including the pulse rate of the diaphragm pump, the fluid temperature, and the fluid viscosity were maintained. Eight VADs were tested with testing periods of 2 years, during which they were continuously in operation. The VAD performance factors, including the power consumption and the vibration characteristics, were kept almost constant. The long-term durability of the developed VAD was successfully demonstrated.

Development of an Optical Detector of Thrombus Formation on the Pivot Bearing of a Rotary Blood Pump.

Continuous optical monitoring of thrombus formation in extracorporeal mechanical circulatory support (EMCS) devices will contribute to safe, long-term EMCS. A clinically applicable optical detector must be able to distinguish among the optical characteristics of oxygen saturation (SaO2 ), hematocrit (Hct), and thrombus formation. In vitro studies of spectral changes at wavelengths from 400 to 900 nm associated with SaO2 , Hct, and thrombus formed around the top pivot bearing of a Gyro C1E3 pump were conducted. Fresh porcine blood anticoagulated with sodium citrate was circulated in a mock circuit using the pump. The SaO2 , Hct, and anticoagulation activity were altered using an oxygenator, autologous plasma, and calcium chlorite injection, respectively. Light from a xenon lamp was guided by an incident fiber perpendicularly fixed on the top bearing. This light was scattered by blood pooled between the male and female pivots. The detection fiber was perpendicularly fixed against the incident fiber, and the side-scattered light was detected and guided to a spectrophotometer. As a result, light at two different wavelengths, 420 and 810 nm, was identified as suitable for thrombus detection because it was negligibly influenced by SaO2 and was able to detect the optical characteristics of fibrin. The light at these two wavelengths responded more quickly to thrombus formation than the inlet or outlet pressure, and flow rate change. The optical changes showed the changes in Hct around the top pivot bearing, which is caused by the reduction in density of fibrin-trapped red blood cells (RBCs) due to the RBCs being swept away by the surrounding blood flow. The proposed method was also able to detect fibrin production by extracting subtle differences in the optical characteristics between the Hct and thrombus formation.

Effect of Impeller Geometry on Lift-Off Characteristics and Rotational Attitude in a Monopivot Centrifugal Blood Pump.

The effect of the flow path geometry of the impeller on the lift-off and tilt of the rotational axis of the impeller against the hydrodynamic force was investigated in a centrifugal blood pump with an impeller supported by a single-contact pivot bearing. Four types of impeller were compared: the FR model with the flow path having both front and rear cutouts on the tip, the F model with the flow path having only a front cutout, the R model with only a rear cutout, and the N model with a straight flow path. First, the axial thrust and the movement about the pivot point, which was loaded on the surface of the impeller, were calculated using computational fluid dynamics (CFD) analysis. Next, the lift-off point and the tilt of the rotational axis of the impeller were measured experimentally. The CFD analysis showed that the axial thrust increased gently in the FR and R models as the flow rate increased, whereas it increased drastically in the F and N models. This difference in axial thrust was likely from the higher pressure caused by the smaller circumferential velocity in the gap between the top surface of the impeller and the casing in the FR and R models than in the F and N models, which was caused by the rear cutout. These results corresponded with the experimental results showing that the impellers lifted off in the F and N models as the flow rate increased, whereas it did not in the FR and R models. Conversely, the movement about the pivot point increased in the direction opposite the side with the pump outlet as the flow rate increased. However, the tilt of the rotational axis of the impeller, which oriented away from the pump outlet, was less than 0.8° in any model under any conditions, and was considered to negligibly affect the rotational attitude of the impeller. These results confirm that a rear cutout prevents lift-off of the impeller caused by a decrease in the axial thrust.

Plasma Skimming in a Spiral Groove Bearing of a Centrifugal Blood Pump.

Plasma skimming is a phenomenon in which discharge hematocrit is lower than feed hematocrit in microvessels. Plasma skimming has been investigated at a bearing gap in a spiral groove bearing (SGB), as this has the potential to prevent hemolysis in the SGB of a blood pump. However, it is not clear whether plasma skimming occurs in a blood pump with the SGB, because the hematocrit has not been obtained. The purpose of this study is to verify plasma skimming in an SGB of a centrifugal blood pump by developing a hematocrit measurement method in an SGB. Erythrocyte observation using a high-speed microscope and a bearing gap measurement using a laser confocal displacement meter was performed five times. In these tests, bovine blood as a working fluid was diluted with autologous plasma to adjust the hematocrit to 1.0%. A resistor was adjusted to achieve a pressure head of 100 mm Hg and a flow rate of 5.0 L/min at a rotational speed of 2800 rpm. Hematocrit on the ridge region in the SGB was measured using an image analysis based on motion image of erythrocytes, mean corpuscular volume, the measured bearing gap, and a cross-sectional area of erythrocyte. Mean hematocrit on the ridge region in the SGB was linearly reduced from 0.97 to 0.07% with the decreasing mean bearing gap from 38 to 21 µm when the rotational speed was changed from 2250 to 3000 rpm. A maximum plasma skimming efficiency of 93% was obtained with a gap of 21 µm. In conclusion, we succeeded in measuring the hematocrit on the ridge region in the SGB of the blood pump. Hematocrit decreased on the ridge region in the SGB and plasma skimming occurred with a bearing gap of less than 30 µm in the hydrodynamically levitated centrifugal blood pump.

Optical aggregometry of red blood cells associated with the blood-clotting reaction in extracorporeal circulation support.

The aggregability of red blood cell (RBCs) is associated with the contribution of plasma proteins, such as fibrinogen and lipoproteids, to blood-clotting. Hence, we hypothesized that RBC aggregability reflects the blood-clotting reaction. A noninvasive optical monitoring method to measure RBC aggregability for the assessment of blood-clotting stage during mechanical circulatory support was developed. An in vitro thrombogenic test was conducted with a rotary blood pump using heparinized fresh porcine blood. Near-infrared laser light at a wavelength of 785 nm was guided by an optical fiber. The fibers for detecting incident, forward-, and backward-scattered light were fixed on the circuit tubing with an inner diameter of 1/4 inch. Because there is substantial RBC aggregation at low shear flow rates, a pulsatile flow was generated by controlling the pump rotational speed. The flow rate was changed from 0 to 8.5 L/min at a period of 40 s. The intensities of forward- and backward-scattered light changed dramatically when the flow stopped. The aggregability was evaluated by the increase ratio of the transmitted light intensity from the flow stopping in the low-flow condition. The experiment started when the anticoagulation was stopped by the addition of protamine into the circulating blood. Reduction in RBC aggregability was associated with a decrease in the amount of fibrinogen and the number of platelets. Continuous, noninvasive monitoring of thrombosis risk is possible using optical measurements combining pulsatile flow control of a rotary blood pump. RBC aggregometry is a potential label-free method for evaluating blood-clotting risk.

Properties of a monopivot centrifugal blood pump manufactured by 3D printing.

An impeller the same geometry as the impeller of a commercial monopivot cardiopulmonary bypass pump was manufactured using 3D printing. The 3D-printed impeller was integrated into the pump casing of the commercially available pump to form a 3D-printed pump model. The surface roughness of the impeller, the hydraulic performance, the axial displacement of the rotating impeller, and the hemolytic properties of the 3D-printed model were measured and compared with those of the commercially available model. Although the surface roughness of the 3D-printed model was significantly larger than that of the commercially available model, the hydraulic performance of the two models almost coincided. The hemolysis level of the 3D-printed model roughly coincided with that of the commercially available model under low-pressure head conditions, but increased greatly under high-pressure head conditions, as a result of the narrow gap between the rotating impeller and the pump casing. The gap became narrow under high-pressure head conditions, because the axial thrust applied to the impeller increased with increasing impeller rotational speed. Moreover, the axial displacement of the rotating impeller was twice that of the commercially available model, confirming that the elastic deformation of the 3D-printed impeller was larger than that of the commercially available impeller. These results suggest that trial models manufactured by 3D printing can reproduce the hydraulic performance of the commercial product. However, both the surface roughness and the deformation of the trial models must be considered to precisely evaluate the hemolytic properties of the model.

Real-Time Observation of Thrombus Growth Process in an Impeller of a Hydrodynamically Levitated Centrifugal Blood Pump by Near-Infrared Hyperspectral Imaging.

Understanding the thrombus formation in cardiovascular devices such as rotary blood pumps is the most important issue in developing more hemocompatible devices. The objective of this study was to develop a hyperspectral imaging (HSI) method to visualize the thrombus growth process within a rotary blood pump and investigate the optical properties of the thrombus. An in vitro thrombogenic test was conducted using fresh porcine blood and a specially designed hydrodynamically levitated centrifugal blood pump with a transparent bottom. The pump rotating at 3000 rpm circulated the blood at 1.0 L/min. The bottom surface of the pump was illuminated with white light pulsed at the same frequency as the pump rotation, and the backward-scattered light was imaged using the HSI system. Using stroboscopic HSI and an image construction algorithm, dynamic spectral imaging at wavelengths ranging from 608 to 752 nm within the rotating pump was achieved. After completing the experiment, we collected the red thrombus formed in the pump impeller and quantified the thrombus hemoglobin concentration (Hbthrombus ). The spectrum changed around the center of the impeller, and the area of change expanded toward the impeller flow path. The shape corresponded approximately to the shape of the thrombus. The spectrum change indicated that the light scattering derived from red blood cells decreased. The Hbthrombus was 4.7 ± 1.3 g/dL versus a total hemoglobin of 13 ± 0.87 g/dL. The study revealed that Hbthrombus was reduced by the surrounding blood flow.

Evaluation of a Spiral Groove Geometry for Improvement of Hemolysis Level in a Hydrodynamically Levitated Centrifugal Blood Pump.

The purpose of this study is to evaluate a spiral groove geometry for a thrust bearing to improve the hemolysis level in a hydrodynamically levitated centrifugal blood pump. We compared three geometric models: (i) the groove width is the same as the ridge width at any given polar coordinate (conventional model); (ii) the groove width contracts inward from 9.7 to 0.5 mm (contraction model); and (iii) the groove width expands inward from 0.5 to 4.2 mm (expansion model). To evaluate the hemolysis level, an impeller levitation performance test and in vitro hemolysis test were conducted using a mock circulation loop. In these tests, the driving conditions were set at a pressure head of 200 mm Hg and a flow rate of 4.0 L/min. As a result of the impeller levitation performance test, the bottom bearing gaps of the contraction and conventional models were 88 and 25 µm, respectively. The impeller of the expansion model touched the bottom housing. In the hemolysis test, the relative normalized index of hemolysis (NIH) ratios of the contraction model in comparison with BPX-80 and HPM-15 were 0.6 and 0.9, respectively. In contrast, the relative NIH ratios of the conventional model in comparison with BPX-80 and HPM-15 were 9.6 and 13.7, respectively. We confirmed that the contraction model achieved a large bearing gap and improved the hemolysis level in a hydrodynamically levitated centrifugal blood pump.

Relative permittivity measurement during the thrombus formation process using the dielectric relaxation method for various hematocrit values.

The relative permittivity ε' and the dielectric loss ε″ for various hematocrit values H for static bovine blood condition have been measured using the dielectric relaxation method to detect thrombosis in real time. The suitable measurement frequency f m ranged within 60 kHz to 1 MHz, and the relaxation frequency of red blood cells (RBCs) f rc was observed to be 2 MHz. In the f m, the temporal change of normalized ε' exhibited a minimum (called as bottom point). The bottom point was observed to be exponentially shortened as H increased. This characteristic of the ε'* minimum is discussed from three viewpoints: during fibrin formation, direct thrombus formation, and rouleaux formation processes. ε'* during the fibrin formation process decreased over time, irrespective of f. However, ε'* in f m during the direct thrombus formation process and during the aggregation formation process increased immediately and rapidly over time. Therefore, the ε'* bottom point in f m might be the indication of micrometer-scale thrombus formation by RBC aggregation due to fibrin formation.

Prediction of heterotrimeric protein complexes by two-phase learning using neighboring kernels.

BACKGROUND: Protein complexes play important roles in biological systems such as gene regulatory networks and metabolic pathways. Most methods for predicting protein complexes try to find protein complexes with size more than three. It, however, is known that protein complexes with smaller sizes occupy a large part of whole complexes for several species. In our previous work, we developed a method with several feature space mappings and the domain composition kernel for prediction of heterodimeric protein complexes, which outperforms existing methods. RESULTS: We propose methods for prediction of heterotrimeric protein complexes by extending techniques in the previous work on the basis of the idea that most heterotrimeric protein complexes are not likely to share the same protein with each other. We make use of the discriminant function in support vector machines (SVMs), and design novel feature space mappings for the second phase. As the second classifier, we examine SVMs and relevance vector machines (RVMs). We perform 10-fold cross-validation computational experiments. The results suggest that our proposed two-phase methods and SVM with the extended features outperform the existing method NWE, which was reported to outperform other existing methods such as MCL, MCODE, DPClus, CMC, COACH, RRW, and PPSampler for prediction of heterotrimeric protein complexes. CONCLUSIONS: We propose two-phase prediction methods with the extended features, the domain composition kernel, SVMs and RVMs. The two-phase method with the extended features and the domain composition kernel using SVM as the second classifier is particularly useful for prediction of heterotrimeric protein complexes.

Feasibility of the optical imaging of thrombus formation in a rotary blood pump by near-infrared light.

Blood coagulation is one of the primary concerns when using mechanical circulatory support devices such as blood pumps. Noninvasive detection and imaging of thrombus formation is useful not only for the development of more hemocompatible devices but also for the management of blood coagulation to avoid risk of infarction. The objective of this study is to investigate the use of near-infrared light for imaging of thrombus formation in a rotary blood pump. The optical properties of a thrombus at wavelengths ranging from 600 to 750 nm were analyzed using a hyperspectral imaging (HSI) system. A specially designed hydrodynamically levitated centrifugal blood pump with a visible bottom area was used. In vitro antithrombogenic testing was conducted five times with the pump using bovine whole blood in which the activated blood clotting time was adjusted to 200 s prior to the experiment. Two halogen lights were used for the light sources. The forward scattering through the pump and backward scattering on the pump bottom area were imaged using the HSI system. HSI showed an increase in forward scattering at wavelengths ranging from 670 to 750 nm in the location of thrombus formation. The time at which the thrombus began to form in the impeller rotating at 2780 rpm could be detected. The spectral difference between the whole blood and the thrombus was utilized to image thrombus formation. The results indicate the feasibility of dynamically detecting and imaging thrombus formation in a rotary blood pump.