SciTech #ScienceSunday Digest - 51/2016.
SciTech #ScienceSunday Digest - 51/2016.
Permalink here: http://www.scitechdigest.net/2016/12/writing-dna-macro-quantum-effects-deep.html
Writing DNA, Macro quantum effects, Autonomous drone advances, Deep learning music, Reversing stem cell development, Reversing animal aging, Photovoltaic atomic veins, Noninvasive mind control, Microfluidic blood sensors, Deep learning supercomputing.
1. Writing DNA with Twist
Twist Bioscience has developed a new silicon chip for writing and synthesising DNA sequences, taking the conventional 96 well approach producing one gene to a new architecture that produces 9,600 genes that makes DNA synthesis significantly faster and cheaper https://medium.com/@Hello_Tomorrow/is-dna-the-next-silicon-c88e6e89754d#.v26ntfksi. Combined with rapid, cheap DNA sequencing chips, rapid, cheap DNA synthesis chips will accelerate the transformation of biotechnology, allowing custom DNA sequences to be produced for quickly engineering organisms of interest.
2. Quantum Effects in Macro Materials
A new type of topological insulator (conductor on surface, insulator in bulk) made from bismuth and selenium, happens to slightly rotate and change a beam of terahertz light shone through the material http://www.sciencealert.com/this-new-material-might-show-the-link-between-classical-and-quantum-physics. This quantum effect is typically observed only at atomic scales and never in macro materials; it obeyed the same mathematics and is the first time such a quantum effect has been observed in large topological insulators. It is hoped that the link might allow further probing between quantum and classical mechanics.
3. Advances with Autonomous Drones
First, an optimal reciprocal collision avoidance strategy has been developed to allow large numbers of drones to fly through the same airspace, dynamically avoid colliding, and all while minimising g-forces for the purpose of future passenger-transport drones http://spectrum.ieee.org/automaton/robotics/drones/limiting-jerks-for-comfortable-commuting-by-personal-drone. Second, Amazon demonstrated its first autonomous drone product delivery https://www.amazon.com/b?node=8037720011. Finally, drones are being used for cheap but difficult environmental monitoring applications such as methane monitoring, an area we can expect to rapidly expand into drone monitoring of a great many things http://spectrum.ieee.org/energywise/energy/environment/drones-take-to-the-skies-to-screen-for-methane-emissions.
4. Deep Learning Music Composition
A new deep learning system called DeepBach was trained and validated against music composed by the composer Bach, and is able to produce new music in the same style as Bach to the extent of convincing humans about 50% of the time that they were actually written by Bach https://www.technologyreview.com/s/603137/deep-learning-machine-listens-to-bach-then-writes-its-own-music-in-the-same-style/. Interestingly the same humans only picked 75% of the compositions actually written by Bach. This marks another big step on the way of machines producing creative outputs and artworks, suggesting that future deep learning systems might produce novel music (or other works) in any particular artist’s style given some general starting parameters.
5. Reversing Human Embryonic Stem Cell Development
A mixture of three different chemical inhibitors has been demonstrated to further wind back the developmental clock of human embryonic stem cells, finally achieving the same long-hoped-for flexibility that researchers have enjoyed with embryonic stem cells from mice https://www.eurekalert.org/pub_releases/2016-12/jhm-rtb121416.php. These stem cells are now much easier to keep alive and the technique successfully reset 25 human stem cell lines, showed more malleable gene expression profiles, avoided abnormal DNA changes sometimes characterised by other techniques, and could be subsequently differentiated into vascular or neural cell types (for example) at double or triple the frequencies of conventional human embryonic stem cells. Meanwhile stem cells are being used to create and study amniotic sac formation http://ns.umich.edu/new/releases/24415-how-does-the-amniotic-sac-form-u-m-team-uses-stem-cells-to-study-earliest-stages.
6. Reversing Aging in Animals
In related work using conventional reprogramming techniques with four factors that turn cells into induced pluripotent stem cells, researchers demonstrated that administering these factors for short durations had rejuvenating, anti-aging effects https://www.salk.edu/news-release/turning-back-time-salk-scientists-reverse-signs-aging/. Skin cells showed reversal of aging hallmarks while remaining skin cells, mice with progeria looked younger with improved organ function and lived 30% longer, while normally aged mice had improved regenerative and healing capacity. As promising as this is it should be approached with caution due to a number of reasons outlined here https://www.fightaging.org/archives/2016/12/temporarily-applying-pluripotency-reprogramming-factors-to-adult-mice/. Meanwhile microRNA levels over time appear to correlate well with life span https://www.fightaging.org/archives/2016/12/microrna-differences-across-the-course-of-aging-correlate-with-life-span/.
7. Atomic Veins Boost Photovoltaic Performance
Adding a network of linear atomic defects by removing atoms on two-dimensional material surfaces such as molybdenum diselenide creates the equivalent of atom-thick wires that can channel electrons and light http://phys.org/news/2016-12-lines-atoms-thin-electronic-materials.html. Early stage research but offering promising avenues to boost photovoltaic performance and explore new properties on these surfaces that influence electrical and optical performance and both semi- and super-conductivity.
8. Noninvasive Mind Control of Robotic Hands
A new 64 electrode EEG system allows people to operate a robotic arm to reach and grasp objects using just their thoughts and without an invasive brain implant https://twin-cities.umn.edu/news-events/umn-research-shows-people-can-control-robotic-arm-their-minds. Studies with human volunteers required them to devote time with the system to learn to imagine moving their own arm, and the robotic arm, without actually moving their arm. In related news a soft prosthetic hand utilises stretchable optical waveguides to detect curvature, elongation, and force and designed to give robots and prosthetics a much better sense of touch http://mediarelations.cornell.edu/2016/12/12/new-robot-has-a-human-touch/.
9. Sensors & Microfluidics for Real-time Blood Monitoring
A microfluidic biosensor chip uses gold electrodes patterned with DNA aptamers to measure molecules of interest in real-time, significantly boosting the accuracy and frequency of measurements and solving other problems that an earlier prototype chip possessed http://spectrum.ieee.org/the-human-os/biomedical/devices/sensor-system-offers-realtime-control-of-drug-levels-in-blood. In one demonstration the concentration of a chemotherapy drug was monitored in rabbits in order to continuously dose the animal with precise amounts of the drug to maintain optimal therapeutic effect while minimising side effects. Different DNA aptamers can be engineered to capture just about any molecule (or combination) of interest, so this is a very interesting platform. In related news another microfluidic chip rapidly detects metastatic cancers cells in drops of blood https://www.wpi.edu/news/wpi-researchers-build-%E2%80%9Cliquid-biopsy%E2%80%9D-chip-detects-metastatic-cancer-cells-drop-blood.
10. Big & Small Supercomputing Initiatives
First, Cray announced the results of a deep learning supercomputing collaboration with Microsoft and the Swiss National Supercomputing Centre that runs larger deep learning models and significantly accelerating the deep learning training process, obtaining results in hours that previously might have taken weeks or months http://investors.cray.com/phoenix.zhtml?c=98390&p=irol-newsArticle&ID=2228098. Second, Nvidia’s DGX-1 supercomputer is a complete dedicated package for machine learning, the size of a briefcase and costing $129,000 that seems to be producing decent advances for ever-more customers https://www.technologyreview.com/s/603075/the-pint-sized-supercomputer-that-companies-are-scrambling-to-get/.
SciTech Tip Jar: http://www.scitechdigest.net/p/donate.html
Permalink here: http://www.scitechdigest.net/2016/12/writing-dna-macro-quantum-effects-deep.html
Writing DNA, Macro quantum effects, Autonomous drone advances, Deep learning music, Reversing stem cell development, Reversing animal aging, Photovoltaic atomic veins, Noninvasive mind control, Microfluidic blood sensors, Deep learning supercomputing.
1. Writing DNA with Twist
Twist Bioscience has developed a new silicon chip for writing and synthesising DNA sequences, taking the conventional 96 well approach producing one gene to a new architecture that produces 9,600 genes that makes DNA synthesis significantly faster and cheaper https://medium.com/@Hello_Tomorrow/is-dna-the-next-silicon-c88e6e89754d#.v26ntfksi. Combined with rapid, cheap DNA sequencing chips, rapid, cheap DNA synthesis chips will accelerate the transformation of biotechnology, allowing custom DNA sequences to be produced for quickly engineering organisms of interest.
2. Quantum Effects in Macro Materials
A new type of topological insulator (conductor on surface, insulator in bulk) made from bismuth and selenium, happens to slightly rotate and change a beam of terahertz light shone through the material http://www.sciencealert.com/this-new-material-might-show-the-link-between-classical-and-quantum-physics. This quantum effect is typically observed only at atomic scales and never in macro materials; it obeyed the same mathematics and is the first time such a quantum effect has been observed in large topological insulators. It is hoped that the link might allow further probing between quantum and classical mechanics.
3. Advances with Autonomous Drones
First, an optimal reciprocal collision avoidance strategy has been developed to allow large numbers of drones to fly through the same airspace, dynamically avoid colliding, and all while minimising g-forces for the purpose of future passenger-transport drones http://spectrum.ieee.org/automaton/robotics/drones/limiting-jerks-for-comfortable-commuting-by-personal-drone. Second, Amazon demonstrated its first autonomous drone product delivery https://www.amazon.com/b?node=8037720011. Finally, drones are being used for cheap but difficult environmental monitoring applications such as methane monitoring, an area we can expect to rapidly expand into drone monitoring of a great many things http://spectrum.ieee.org/energywise/energy/environment/drones-take-to-the-skies-to-screen-for-methane-emissions.
4. Deep Learning Music Composition
A new deep learning system called DeepBach was trained and validated against music composed by the composer Bach, and is able to produce new music in the same style as Bach to the extent of convincing humans about 50% of the time that they were actually written by Bach https://www.technologyreview.com/s/603137/deep-learning-machine-listens-to-bach-then-writes-its-own-music-in-the-same-style/. Interestingly the same humans only picked 75% of the compositions actually written by Bach. This marks another big step on the way of machines producing creative outputs and artworks, suggesting that future deep learning systems might produce novel music (or other works) in any particular artist’s style given some general starting parameters.
5. Reversing Human Embryonic Stem Cell Development
A mixture of three different chemical inhibitors has been demonstrated to further wind back the developmental clock of human embryonic stem cells, finally achieving the same long-hoped-for flexibility that researchers have enjoyed with embryonic stem cells from mice https://www.eurekalert.org/pub_releases/2016-12/jhm-rtb121416.php. These stem cells are now much easier to keep alive and the technique successfully reset 25 human stem cell lines, showed more malleable gene expression profiles, avoided abnormal DNA changes sometimes characterised by other techniques, and could be subsequently differentiated into vascular or neural cell types (for example) at double or triple the frequencies of conventional human embryonic stem cells. Meanwhile stem cells are being used to create and study amniotic sac formation http://ns.umich.edu/new/releases/24415-how-does-the-amniotic-sac-form-u-m-team-uses-stem-cells-to-study-earliest-stages.
6. Reversing Aging in Animals
In related work using conventional reprogramming techniques with four factors that turn cells into induced pluripotent stem cells, researchers demonstrated that administering these factors for short durations had rejuvenating, anti-aging effects https://www.salk.edu/news-release/turning-back-time-salk-scientists-reverse-signs-aging/. Skin cells showed reversal of aging hallmarks while remaining skin cells, mice with progeria looked younger with improved organ function and lived 30% longer, while normally aged mice had improved regenerative and healing capacity. As promising as this is it should be approached with caution due to a number of reasons outlined here https://www.fightaging.org/archives/2016/12/temporarily-applying-pluripotency-reprogramming-factors-to-adult-mice/. Meanwhile microRNA levels over time appear to correlate well with life span https://www.fightaging.org/archives/2016/12/microrna-differences-across-the-course-of-aging-correlate-with-life-span/.
7. Atomic Veins Boost Photovoltaic Performance
Adding a network of linear atomic defects by removing atoms on two-dimensional material surfaces such as molybdenum diselenide creates the equivalent of atom-thick wires that can channel electrons and light http://phys.org/news/2016-12-lines-atoms-thin-electronic-materials.html. Early stage research but offering promising avenues to boost photovoltaic performance and explore new properties on these surfaces that influence electrical and optical performance and both semi- and super-conductivity.
8. Noninvasive Mind Control of Robotic Hands
A new 64 electrode EEG system allows people to operate a robotic arm to reach and grasp objects using just their thoughts and without an invasive brain implant https://twin-cities.umn.edu/news-events/umn-research-shows-people-can-control-robotic-arm-their-minds. Studies with human volunteers required them to devote time with the system to learn to imagine moving their own arm, and the robotic arm, without actually moving their arm. In related news a soft prosthetic hand utilises stretchable optical waveguides to detect curvature, elongation, and force and designed to give robots and prosthetics a much better sense of touch http://mediarelations.cornell.edu/2016/12/12/new-robot-has-a-human-touch/.
9. Sensors & Microfluidics for Real-time Blood Monitoring
A microfluidic biosensor chip uses gold electrodes patterned with DNA aptamers to measure molecules of interest in real-time, significantly boosting the accuracy and frequency of measurements and solving other problems that an earlier prototype chip possessed http://spectrum.ieee.org/the-human-os/biomedical/devices/sensor-system-offers-realtime-control-of-drug-levels-in-blood. In one demonstration the concentration of a chemotherapy drug was monitored in rabbits in order to continuously dose the animal with precise amounts of the drug to maintain optimal therapeutic effect while minimising side effects. Different DNA aptamers can be engineered to capture just about any molecule (or combination) of interest, so this is a very interesting platform. In related news another microfluidic chip rapidly detects metastatic cancers cells in drops of blood https://www.wpi.edu/news/wpi-researchers-build-%E2%80%9Cliquid-biopsy%E2%80%9D-chip-detects-metastatic-cancer-cells-drop-blood.
10. Big & Small Supercomputing Initiatives
First, Cray announced the results of a deep learning supercomputing collaboration with Microsoft and the Swiss National Supercomputing Centre that runs larger deep learning models and significantly accelerating the deep learning training process, obtaining results in hours that previously might have taken weeks or months http://investors.cray.com/phoenix.zhtml?c=98390&p=irol-newsArticle&ID=2228098. Second, Nvidia’s DGX-1 supercomputer is a complete dedicated package for machine learning, the size of a briefcase and costing $129,000 that seems to be producing decent advances for ever-more customers https://www.technologyreview.com/s/603075/the-pint-sized-supercomputer-that-companies-are-scrambling-to-get/.
SciTech Tip Jar: http://www.scitechdigest.net/p/donate.html
Comments
Post a Comment