SIMON: Open-Source Knowledge Discovery Platform

Data analysis and knowledge discovery has become more and more important in biology and medicine with the increasing complexity of biological datasets, but the necessarily sophisticated programming skills and in-depth understanding of algorithms needed pose barriers to most biologists and clinicians to perform such research. We have developed a modular open-source software, SIMON, to facilitate the application of 180+ state-of-the-art machine-learning algorithms to high-dimensional biomedical data. With an easy-to-use graphical user interface, standardized pipelines, and automated approach for machine learning and other statistical analysis methods, SIMON helps to identify optimal algorithms and provides a resource that empowers non-technical and technical researchers to identify crucial patterns in biomedical data.

High-throughput ultra-high-performance liquid chromatography/tandem mass spectrometry quantitation of insulin-like growth factor-I and leucine-rich alpha-2-glycoprotein in serum as biomarkers of recombinant human growth hormone administration

Insulin-like growth factor-I (IGF-I) is a known biomarker of recombinant human growth hormone (rhGH) abuse, and is also used clinically to confirm acromegaly. The protein leucine-rich alpha-2-glycoprotein (LRG) was recently identified as a putative biomarker of rhGH administration. The combination of an ACN depletion method and a 5-min ultra-high-performance liquid chromatography/tandem mass spectrometry (uHPLC/MS/MS)-based selected reaction monitoring (SRM) assay detected both IGF-I and LRG at endogenous concentrations. Four eight-point standard addition curves of IGF-I (16-2000 ng/mL) demonstrated good linearity (r(2) = 0.9991 and coefficients of variance (CVs) <13%). Serum samples from two rhGH administrations were extracted and their uHPLC/MS/MS-derived IGF-I concentrations correlated well against immunochemistry-derived values. Combining IGF-I and LRG data improved the separation of treated and placebo states compared with IGF-I alone, further strengthening the hypothesis that LRG is a biomarker of rhGH administration. Artificial neural networks (ANNs) analysis of the LRG and IGF-I data demonstrated an improved model over that developed using IGF-I alone, with a predictive accuracy of 97%, specificity of 96% and sensitivity of 100%. Receiver operator characteristic (ROC) analysis gave an AUC value of 0.98. This study demonstrates the first large scale and high throughput uHPLC/MS/MS-based quantitation of a medium abundance protein (IGF-I) in human serum. Furthermore, the data we have presented for the quantitative analysis of IGF-I suggest that, in this case, monitoring a single SRM transition to a trypsin peptide surrogate is a valid approach to protein quantitation by LC/MS/MS.

A variable inertial system for measuring the contractile properties of human muscle

PURPOSE

A flywheel system of variable inertia is described for inferring the mechanical properties of human muscle during a single explosive movement.

METHODS

The system consists of a lightweight aluminum disk mounted on a shaft onto which a driving cog is mounted. The inertia of the system can be varied from 0.024 to 0.69 kg.m(2) by attaching semicircular steel plates to the disk. A rotary encoder detects displacement of the wheel with a resolution of 1 degrees. Digital signals from the encoder are collected using an A/D converter interfaced to a PC. The data are then processed for the calculation of torque, velocity, power, work done, and acceleration. The mechanical properties of the muscles employed are inferred from calculations of flywheel displacement, time, and force. In addition, a pretension release mechanism can be incorporated into the system to allow isometric force to be developed before movement. This can increase power generation at the low inertias where the time of contraction is typically less than 200 ms. Seven subjects were test-retested using the device. Measures of both average and peak power were made.

RESULTS

When mounted in the apparatus described by Bassey and Short, the maximum values for peak and average power were on average 965 +/- 103 and 448 +/- 47 W, respectively. Upon retesting, these results were found to be reliable (cv = 3.3% and 3.0%, respectively).

CONCLUSIONS

The inertial system described has been shown to have validity in reproducibility and provided a suitable method of determining a number of muscle output properties during short-term single exertions. This tool could prove useful in a research or clinical setting and may also prove useful as a training device as it negates the need for a strain gauge or goniometer attachment.

High expression of MHC I in the tibialis anterior muscle of a paraplegic patient

A long-term paraplegic man presented exclusively (>99%) myosin heavy chain I (MHC I) in the tibialis anterior muscle (TA). This was coupled to a slow speed of contraction, a high resistance to fatigue, and a rapid resynthesis of phosphocreatine after an electrically evoked fatiguing contraction when compared with the TA muscles of 9 other paraplegic individuals. In contrast, the MHC composition of his vastus lateralis, gastrocnemius, and soleus muscles was that expected of a muscle from a spinal cord injured individual. This information may be of clinical importance in terms of the expected morphological and functional adaptations of skeletal muscle to different types of electrical stimulation therapy.

Knee extensor strength, activation, and size in very elderly people following strength training

Muscle strength, activation, and size were studied in 11 very elderly subjects (8 women and 3 men; age range, 85-97 years) who completed 12 weeks of strength training of the knee extensor muscles. Training increased the maximum amount of weight that could be lifted once (134%; P < 0.05) and maximum voluntary isometric strength, measured as both force recorded at the ankle with the knee flexed 90 degrees (17%, ns) and as torque with the knee flexed 60 degrees (37%; P < 0.05). Anatomical lean quadriceps cross-sectional area (LCSA) measured at midthigh using magnetic resonance imaging increased from 27.5 +/- 9.6 cm2 to 30.2 +/- 10.0 cm2 (9.8%; P < 0. 05) after training. Both before and after training, isometric strength was closely related to LCSA, but training resulted in no significant change in muscle force per unit area of quadriceps muscle. Using the twitch interpolation technique, muscle activation during a maximal voluntary isometric contraction was shown to be incomplete in all subjects before training (ranging from 69% to 93%) and was not significantly increased after training. An increase in skeletal muscle mass may have important functional and metabolic benefits for very elderly people.

Sprint training, in vitro and in vivo muscle function, and myosin heavy chain expression

Sprint training represents the condition in which increases in muscle shortening speed, as well as in strength, might play a significant role in improving power generation. This study therefore aimed to determine the effects of sprint training on 1) the coupling between myosin heavy chain (MHC) isoform expression and function in single fibers, 2) the distribution of MHC isoforms across a whole muscle, and 3) in vivo muscle function. Seven young male subjects completed 6 wk of training (3-s sprints) on a cycle ergometer. Training was without effect on maximum shortening velocity in single fibers or in the relative distribution of MHC isoforms in either the soleus or the vastus lateralis muscles. Electrically evoked and voluntary isometric torque generation increased (P < 0.05) after training in both the plantar flexors (+8% at 50 Hz and +16% maximal voluntary contraction) and knee extensors (+8% at 50 Hz and +7% maximal voluntary contraction). With the shortening potential of the muscles apparently unchanged, the increased strength of the major lower limb muscles is likely to have contributed to the 7% increase (P < 0.05) in peak pedal frequency during cycling.

Whole-muscle and single-fibre contractile properties and myosin heavy chain isoforms in humans

The contractile characteristics of three human muscle groups (triceps surae, quadriceps femoris and triceps brachii) of seven young male subjects were examined. The contractile properties were determined from electrically evoked isometric responses and compared with fibre type composition determined from needle biopsy samples. Fibre types were identified using myosin heavy chain (MHC) isoforms as molecular markers with gel electrophoresis (SDS-PAGE) and histochemical ATPase staining. Four contractile parameters (twitch time to peak torque, the maximal rate of torque development, frequency response and fatiguability) were found to be related to fibre type composition. From the biopsy samples, single muscle fibres were isolated and chemically skinned. Isometric tension (Po) unloaded shortening velocity (Vo) and rate of tension rise (dP/dt) were determined. Each fibre was classified on the basis of its MHC isoform composition determined by SDS-PAGE. Fibres belonging to the same type showed identical contractile parameters regardless of the muscle of origin, except minor differences in Po of the fast fibres and dP/dt of slow fibres. The results are in favour of the conclusion that fibre type composition, determined using MHC isoforms as markers, is the major determinant of the diversity of contractile properties among human muscle groups.

Skeletal muscle contractile characteristics and fatigue resistance in patients with chronic heart failure

Whole muscle contractile characteristics and fatigue resistance were studied in male patients with chronic heart failure (n = 6) and in healthy control subjects (n = 6). Maximum voluntary isometric strength in the major muscle groups of leg (plantar flexors and knee extensors) and arm (elbow extensors and elbow flexors), was found to be similar for both groups of subjects. However, a faster isometric twitch time course was observed in the plantar flexor and knee extensor muscles of heart failure chronic patients. The poor resistance to fatigue in the knee extensors of chronic heart failure patients was confirmed in the present study, but using twitch interpolation this was shown not to be due to poor activation. The plantar flexors of chronic heart failure patients also showed a tendency to be less resistant to fatigue, even when the muscle was activated by direct electrical stimulation. The present study shows that independent of muscle strength, patients with chronic heart failure may possess muscles that are faster to contract and less resistant to fatigue. However, it seems this increased fatigability is not due to poor muscle activation.

Electrically evoked torque-velocity characteristics and isomyosin composition of the triceps surae in young and elderly men

The electrically evoked isokinetic torque-velocity relationship of the triceps surae of eight elderly and four non-trained young men was examined in relation to the isomyosin composition of the soleus and the gastrocnemius muscles, determined under non-denaturing conditions using pyrophosphate gel electrophoresis. The angle specific torque-velocity properties of the triceps surae were measured using maximal percutaneous electrical stimulation at 50 Hz and a release technique. The elderly subjects generated significantly (P < 0.05) less absolute torque at all angular velocities when compared with the young subjects. When the isokinetic data were normalized to the isometric torque, the lower normalized torques generated by the elderly subjects were not statistically different from the young. The total fast isomyosin (FM) content of the soleus and gastrocnemius in the elderly subjects was 22 +/- 13 and 35+/- 18%, respectively. This compared with 29 +/- 8 (n.s) and 44 +/- 8% (n.s.) in the young subjects. When the gastrocnemius and soleus muscles were given an equal weighting and considered together to represent the whole triceps surae, the normalized torque at the fixed angular velocity of 5 rads s-1 was significantly associated with %FM (r = 0.90, P < 0.01), and the isomyosin bands %FM1 (r = 0.90, P < 0.01) and %FM2 (r = 0.93, P < 0.001) when only the elderly subjects were considered. No relationships were observed between contractile characteristics and contractile protein profile when only the young subjects were considered. This was despite the inclusion of a further two sprint and three endurance trained athletes to increase the range of contractile characteristics and differences in muscle composition.

The relationship between the pressor response to involuntary isometric exercise and the contractile protein profile of the active muscle in man

The present investigation examined the relationship between the pressor response during electrically evoked isometric ankle plantar flexion and the contractile protein profile of the active muscle in seven young men [mean (SD) age, 26 (6) years] and five older men [70 (4) years]. Muscle biopsy samples were taken from lateral gastrocnemius (LG) and soleus (SOL) of each subject. These were analysed for isomyosin composition using non-denaturing pyrophosphate polyacrylamide gel electrophoresis. The degree of association was examined between the cardiovascular changes and the fast isomyosin content of LG and SOL individually and in combination (SOL/LG). In the total subject group there was no association between the heart rate response or the change in systolic blood pressure (BP) and the fast isomyosin composition. However, the change in diastolic BP was significantly associated with the fast isomyosin composition of SOL/LG (delta diastolicBP = 0.31 + 0.045% FM SOL/LG, r = 0.65, P = 0.029). These findings suggest that the magnitude of the peripheral reflex mediated pressor response to isometric exercise and the fast isomyosin content of the active muscle are related.

A comparison of voluntary and electrically evoked isokinetic plantar flexor torque in males

The angle-specific isokinetic torque- and power-velocity relationships of the triceps surae were examined in ten male sprint athletes aged [mean (SD)] 22.4 (3.2) years, ten non-trained adult men aged 27.4 (4.8) years and six elderly male subjects aged 68.5 (2.4) years. Normal voluntary contractions were compared with those obtained using maximal tetanic stimulation and a release technique which standardised the level of muscle activation during isokinetic contractions. When the isokinetic data was normalized to the maximum isometric torque the stimulated release contractions at 5.18-5.29 rad.s-1 produced significantly (P < 0.05) greater torque than the voluntary no-release contractions at the same angular velocity in each group of subjects. The three subject groups generated their peak power at 3.07 rad.s-1 during the voluntary no release contractions. However, with the stimulated release contractions, power had still not reached a peak at 5.29 rad.s-1, the highest angular velocity that could be tested. It appears that at higher angular velocities the triceps surae is capable of greater torque and power generation when contractions are evoked using a stimulated release technique. It is suggested that the stimulated release technique gives a more complete picture of the torque-velocity characteristics of the contractile component of the triceps surae.

Muscle activation and the isokinetic torque-velocity relationship of the human triceps surae

The influence of muscle activation and the time allowed for torque generation on the angle-specific torque-velocity relationship of the triceps surae was studied during plantar flexion using supramaximal electrical stimulation and a release technique on six male subjects [mean (SD) age 25 (4) years]. Torque-velocity data were obtained under different levels of constant muscle activation by varying the stimulus frequency and the time allowed for isometric torque generation prior to release and isokinetic shortening. To eliminate the effects of the frequency response on absolute torque the isokinetic data were normalized to the maximum isometric torque values at 0.44 rad. There were no significant differences in the normalized torques generated at any angular velocity using stimulus frequencies of 20, 50 or 80 Hz. When the muscle was stimulated at 50 Hz the torques obtained after a 400 ms and 1 s pre-release isometric contraction did not differ significantly. However, with no pre-release contraction significantly less torque was generated at all angular velocities beyond 1.05 rad.s-1 when compared with either the 200, 400 ms or 1 s condition. With a 200 ms pre-release contraction significantly less torque was generated at angular velocities beyond 1.05 rad.s-1 when compared with the 400 ms or 1 s conditions. It would seem that the major factor governing the shape of the torque-velocity curve at a constant level of muscle activation is the time allowed for torque generation.