A Crowdsourcing Approach to Developing and Assessing Prediction Algorithms for AML Prognosis.

Acute Myeloid Leukemia (AML) is a fatal hematological cancer. The genetic abnormalities underlying AML are extremely heterogeneous among patients, making prognosis and treatment selection very difficult. While clinical proteomics data has the potential to improve prognosis accuracy, thus far, the quantitative means to do so have yet to be developed. Here we report the results and insights gained from the DREAM 9 Acute Myeloid Prediction Outcome Prediction Challenge (AML-OPC), a crowdsourcing effort designed to promote the development of quantitative methods for AML prognosis prediction. We identify the most accurate and robust models in predicting patient response to therapy, remission duration, and overall survival. We further investigate patient response to therapy, a clinically actionable prediction, and find that patients that are classified as resistant to therapy are harder to predict than responsive patients across the 31 models submitted to the challenge. The top two performing models, which held a high sensitivity to these patients, substantially utilized the proteomics data to make predictions. Using these models, we also identify which signaling proteins were useful in predicting patient therapeutic response.

Mechanisms underlying the additive and redundant Qrr phenotypes in Vibrio harveyi and Vibrio cholerae.

Vibrio harveyi and Vibrio cholerae regulate their virulence factors according to the local cell-population density in a regulatory system called quorum sensing. Their quorum sensing systems contain a small RNA (sRNA) circuit to regulate expression of a master transcriptional regulator via multiple quorum regulated RNA (Qrr) and a protein chaperon Hfq. Experiments and genetic analysis show that their respective quorum sensing networks are topologically equivalent and have homologous components, yet they respond differently to the same experimental conditions. In particular, V. harveyi Qrr are additive because all of its Qrr are required to maintain wild-type-like repression of its master transcriptional regulator. Conversely, V. cholerae Qrr are redundant because any of its Qrr is sufficient to repress its master transcriptional regulator. Given the striking similarities between their quorum sensing systems, experimentalists have been unable to identify conclusively the mechanisms behind these phenotypic differences. Nevertheless, the current hypothesis in the literature is that dosage compensation is the mechanism underlying redundancy. In this work, we identify the mechanisms underlying Qrr redundancy using a detailed mathematical model of the V. harveyi and V. cholerae sRNA circuits. We show that there are exactly two different cases underlying Qrr redundancy and that dosage compensation is unnecessary and insufficient to explain Qrr redundancy. Although V. harveyi Qrr are additive when the perturbations in Qrr are large, we predict that V. harveyi and V. cholerae Qrr are redundant when the perturbations in Qrr are small. We argue that the additive and redundant Qrr phenotypes can emerge from parametric differences in the sRNA circuit. In particular, we find that the affinity of Qrr and its expression relative to the master transcriptional regulator determine the level of redundancy in V. harveyi and V. cholerae. Furthermore, the additive and redundant Qrr phenotypes reflect differences in the concentration of Hfq-Qrr in V. harveyi and V. cholerae. We use our model to test the dosage compensation hypothesis and show that decreasing the expression of qrr, rather than removing dosage compensation, abolishes Qrr redundancy in V. cholerae. Further experimentation is needed to test our results and both Qrr redundancy hypotheses.

A framework for modeling of mechano-electrical feedback mechanisms of cardiac myocytes and tissues.

The term cardiac mechano-electrical feedback precis the various phenomena related to modulation of electrophysiology by mechanical deformation of cells and tissues of the heart. The significance of mechano-electrical feedback and the underlying mechanisms are still poorly understood despite intense experimental research. We introduce and discuss a framework for computational modeling and simulation of mechano-electrical feedback at ion channel, cell and tissue level. The framework consists of modules to reconstruct electrical currents through mechano-sensitive ion channels, their effect on myocytes' electrophysiology, strain-modulation of tissue conductivities and electrical conduction in myocyte clusters and myocardium. We applied the framework to study the effect of strain on conduction velocity in papillary muscle. The simulations reconstructed strain-conduction velocity relationships as reported in experimental studies. Furthermore, the computational studies indicated that increased stimulus frequency aggravated the reduction of conduction velocity for larger strains. Mathematical modeling of mechano-electrical feedback will help to integrate experimental data and predict behavior at system level. Computational simulations might give otherwise unavailable insights, particularly with respect to clinical relevance of mechano-electrical feedback.