A new approach to detect and measure echoes in data has been formulated, which is applicable to cases where irregular sampling is concerned.

Proving the feasibility of this approach, required a significant computational effort, which would have taken a normal computer about 10 years to complete. Hive has been instrumental in finishing this project in finite time. The accompanying image shows the results of Monte Carlo simulations that benchmark the ability of the various statistical estimators considered to uncover the input lag (at 100 days). Paper has been submitted to the Astrophysical Journal.

In a first-of-its-kind effort to accurately estimate mass of super-massive black holes, researchers led by Prof. Doron Chelouche, are carrying out nightly photo-metric observations at the Wise Observatory with a follow-up analysis using novel image reduction techniques. This project has far-reaching implications for the formation and evolution of supermassive black holes, the extreme physics in the immediate vicinity of such objects, as well as for the use of quasars as standard cosmological candles. 

The outlined study can only be accomplished using a large computer with many processors and a fast distributed file system. Specifically, a normal computer would be unable to cope with the influx of data from the telescope, and the demanding computational task of analyzing them.

The attached image shows the telescope in the inset, and an example for a light curve (i.e., the light-intensity variation over time) for a particular object, obtained by Dr. Francisco Pozo-Nuñez, which harbors a black hole at its center. Several papers are in preparation.


Tal Yahav from the Privman lab is using the "Queens" of the Hive (high memory servers) to assemble the genome of a new ant species that he himself collected, extracted, and sequenced. The species Cataglyphis drusus is a desert ant native to Israel, which will serve as the model organism for a study in our lab looking for the genetic basis of nestmate recognition in ants.
Tal extracted both DNA and RNA from a single male ant. Genomic DNA sequencing yielded 46 GigaBase (46E9 bases) of sequence data, which Tal assembled into a first draft sequence of the genome. This type of analysis requires very large memory, which is why we purchased the Queens - servers with 768 GB RAM.

Prof. Doron Chelouche and colleagues from Princeton University have detected, for the first time in the local universe, a gigantic gaseous disk revolving around a galaxy. The size of the disk exceeds the visible part of the galaxy by large factor. This discovery was made possible by acquiring unique spectroscopic data using the Hubble Space Telescope.

Studying the physical properties of this disk has made use of Hive's unique computing capabilities in solving thousands of coupled, non-linear differential equations to uncover the physical state of the gas.

The accompanying image shows the galaxy; the recently detected gaseous component, in the form of a titled disk around it, is marked.

Results have recently been published by the Astrophysical Journal.


Pnina Cohen from the Privman lab ran a population genomic analysis pipeline on the Hive to analyze a large-scale sequencing of population samples from the native range (South America) and the introduced range (USA) of two fire ants species.
The data included genomic sequencing (RAD-seq) of 835 ants sampled over many years by our collaborators from Florida, which resulted in 1.8 to 5 million sequences (RAD-tags) per ant.
Using these data, Pnina identified 169,682 polymorphic sites across the fire ant genome. These results are now used for detailed reconstruction of the demographic history of these populations.
This will also allow inference of natural selection pressures on specific genomic regions, indicating genes involved in local adaptation of these invasive ants to different environments. Pnina recently presented these results in a meeting in Helsinki of the International Union for the Study of Social Insects.

Dr. Muhammad Akashi who is working with Prof. Doron Chelouche has been running FLASH code models to simulate the formation and evolution of planetary nebulae formed by aging stars.

Link: http://hivehpc.haifa.ac.il/images/dens3D.png

The above picture shows the density structure of the planetary nebula expanding out from three interacting stars. It's the first time such a calculation has been done with such detail, and what we see is a particular frame in time, for a particular projection of the (three-dimensional) nebula.
For comparison, A real planetary nebula, as seen in space is looking like that:

Source of the picture: http://www.esa.int












Link(source: www.esa.int ): http://www.esa.int/var/esa/storage/images/esa_multimedia/images/2015/07/born-again_planetary_nebula/15533530-1-eng-GB/Born-again_planetary_nebula.jpg

These results have just been presented in a big meeting on the physics of planetary nebulae by Muhammad in China.

Inbal Mermershtian from the Kosloff lab used a computational structural bioinformatics pipeline that requires hundreds of cores working in parallel for accurate atomic-level calculations of protein-protein interactions - analyzing multiple proteins 3D complexes quantitatively.
In this project Inbal focused on the small G-proteins Rabs, comparing their modulation by host GAP proteins (that turn Rab switches “off”) with the modulation by pathogen GAP proteins. She use a computational method developed in the Kosloff lab that applies structure-based electrostatic energy calculations to pinpoint structural determinants that are critical for fine-tuning protein-protein interaction specificity. 
Inbal's analysis uncovers a convergent structural basis for Rab recognition and modulation by eukaryotic and bacterial pathogens, and highlighted key specificity-determining interactions that might be used to target such pathogens with new antibiotics. More generally, this study presents a fascinating case-study of how similar interaction specificity can be achieved by structurally-dissimilar proteins.