Cover photo
Verified name
87,340 followers|16,544,555 views
Have them in circles
87,340 people
Alexander Rauscha's profile photo
manish hirapara's profile photo
Odair Alfredo Gomes's profile photo
Yasmin Pukin's profile photo
Finn Waage Michaelsen's profile photo
Dawn Mason's profile photo
Sahda Haroon's profile photo
rebecca kent's profile photo
Dan Epstein's profile photo
Official page of #ScienceSunday and #ScienceEveryday (brought to you by Robby Bowles, Allison Sekuler, Rajini Rao, Chad Haney, Buddhini Samarasinghe, Aubrey Francisco, and Carissa Braun)
You have found the official page for #ScienceSunday (co-curated by Robby Bowles, Allison Sekuler, Rajini Rao, Chad HaneyBuddhini SamarasingheAubrey Francisco, and +Carissa Braun)! Post anything related to science and tag it with #ScienceSunday, +ScienceSunday, and each curator to ensure we see your post. If you are a photographer, post a science related image and explain why it relates to science to you - doesn't need to be too specific (science is all around us!). If you are not a photographer, simply post anything related to science - drawings, movies, songs, and text are all welcome. Regardless of the type of post, feel free to add your 2 cents into a discussion in the comments. We always have some great posts with amazing images, great science information, and a lot of interesting conversations, and we're looking forward to even more in the weeks to come. If you miss the "Sunday" in #ScienceSunday, feel free to tag with #ScienceEveryday - we try to monitor those posts as well.



Shared publicly  - 
How are the scientific method, free market, and natural selection related?
Read +Sabine Hossenfelder's blog post to find out. I particularly like this paragraph:

In science, the most relevant restriction is that we can’t just randomly generate hypotheses because we wouldn’t be able to test and evaluate them all. This is why science heavily relies on education standards, peer review, and requires new hypotheses to tightly fit into existing knowledge. We also need guidelines for good scientific conduct, reproducibility, and a mechanism to give credits to scientists with successful ideas. Take away any of that and the system wouldn’t work.

I write to try to undo the hype in a new scientific findings where a newspaper has lathered on too much hype. I also moderate the Science on Google+ community. So I often get comments about how we should question everything, that science is about challenging everything. If you don't question everything, e.g., evolution, climate change, etc. then you aren't doing science. Sadly, these comments often come from climate change deniers, believers in pseudoscience or conspiracies.  So I often have to explain that skepticism is fine, however, when you have an extraordinary claim, you need extraordinary evidence. I've written about that before.

Skepticism doesn't equal question all things

I'm not sure many people truly understand the scientific method and IFLS doesn't help with catchy GIFs with no science or attribution.

Science is not about certainty. Science is about finding the most reliable way of thinking, at the present level of knowledge. Science is extremely reliable; it's not certain. In fact, not only it's not certain, but it's the lack of certainty that grounds it. Scientific ideas are credible not because they are sure, but because they are the ones that have survived all the possible past critiques, and they are the most credible because they were put on the table for everybody's criticism. [...]

Image source:


h/t +Filippo Salustri 
22 comments on original post
Chad Haney's profile photoJesse Morgan's profile photonavi maan's profile photo
Nyssh to gain the knowledge
Add a comment...


Shared publicly  - 
Another excellent summary from +Mark Bruce.
Thanks as always.

#ScienceEveryday   #SciSunCH  
SciTech #ScienceSunday Digest - 43/2015.
Permalink here:
Late due to awesome 1,500km road trip with no Internet!

Zero-index metamaterial, pneumatic artificial muscles, DNA looping, thermal energy metamaterials, tensegrity robots, Cognitive algorithms, Smartphone disruption, Tissue building tools, Robot microbial fuel cell, Anti-Aging interventions. 

1. Zero-Index Metamaterial
A new on-chip metamaterial has been developed with a refractive index of zero Zero index metamaterials produce unusual phenomena when light enters them, adopting a uniform infinite wavelength and infinite phase velocity while in the material, and allowing the light to be stretched, twisted, compressed, or turned without losing energy. The material consists of a polymer matrix embedded with silicon pillars and clad in gold film. Such a material opens the door to better entanglement of qubits, novel lasers, lossless on-chip light transmission, optical computer components, and other devices able to manipulate light in new ways. 

2. 3D Printed Pneumatic Artificial Muscles
3D printing via digital mask projection stereolithography has been used to create flexible elastomeric structures with embedded 3D pneumatic actuators that, when attached to a suitable pneumatic system allows the structures to move, bend, and flex in a controllable manner These little tentacles have four degrees of freedom and can sweep through a full range of 180 degree motion in 70 milliseconds. Ultimately the group hopes this proof-of-concept leads to sophisticated biomimetic robotic systems. 

3. Structure, Function, and DNA Looping
This is a good article covering the recent work and growing understanding of the functional structure of DNA loops and still poorly-understood three dimensional maps of the human genome This explores the role of binding and loop proteins, rules for loop formation, the regulatory effect of these large structures, the risk of disease states arising from looping defects even though the DNA sequence of “diseased” genes is not mutated, and new theoretical models for coordinating all of this activity based on an “extrusion complex.” The new models are incredibly accurate at predicting actual physical phenomena in cells and the genome, and yet such an extrusion complex has not yet been identified. Such work should lead to more sophisticated synthetic biology tools. 

4. Metamaterials for Thermal Energy Harvesting
A new metamaterial rectenna is able to effectively harvest thermal waste heat and convert it into direct current This involved detailed simulations of surface topologies and the optimal conditions for capturing and directing surface plasmons, which fed into designing optimal metamaterial surface structures able to harvest such energy. The group awaits experimental validation, which will allow accurate comparisons to competitive technologies such as thermoelectrics. In related news black phosphorous has been shown to have opposite anisotropy for thermal and electric conductivities, with heat and electricity flowing easiest in opposite directions, a key thermoelectric property and a finding that could impact on-chip devices in future

5. New Tensegrity Robots
I’m always on the lookout for new tensegrity robots and structures in general. This new double tetrahedron tensegrity robot is able to autonomously crawl up and along inside tubes and pipes, and might possibly be used as a robust tool to monitor, inspect, repair, and clean different pipes and tube infrastructure. As usual the cable and rod design is robust and versatile and allows a variety of options for incorporating a payload or sensor arrays. Also of interest to robotic designs in future is the fact that the human shoulder joint has a double tetrahedral tensegrity architecture. 

6. Algorithms Replacing Complex Human Tasks
First, the Data Science Machine is a new system designed to not only search for patterns in big datasets but also chooses the feature set to analyse too, and in three data science competitions it outperformed 615 of 906 teams, making predictions 94-96% as accurate in at least two of the competitions, and doing so in 2 - 12 hours rather than the months the human teams used. Second, automatic crowd-counting software can now deduce the populations of large crowds from photos, reaching 530,000 individuals in a demonstration Third, order-of-magnitude speed-ups for some problems are promised by a new general-purpose optimisation algorithm

7. More Smartphone Disruption
A company called EyeNetra is launching a device called the Netra, which is designed to hold a user’s smartphone and be worn over the eyes like a virtual reality display; the user turns dials to interact and align patterns displayed by the device, which is then able to calculate - based on responses - any refractive errors the person has and provide the refractive powers, axis of astigmatism, and pupillary distance required for glasses or contact lens prescription As this develops it should significantly obviate the need for optometrists, reduce costs, and boost convenience for people now able to do this at home; they’re also partnering with VR companies for corrective displays. Meanwhile smartphones can now be used to analyse paper samples for pesticides and possibly other chemicals in future

8. New Tools for Building Tissues
First, another good example of innovation in the 3D bioprinting space to produce structures composed of soft gels inside of and temporarily supported by another block of soft gel that can be dissolved away, and so allowing detailed soft 3D structures to be formed that don’t immediately collapse. Next step is loading with desired cells that grow and colonise the soft 3D bioscaffold. Second, Collymers have been commercially lunched as new self-assembled tissue building blocks, aimed at both research and medical applications, and on the back of successful testing in forming blood vessels and fat- or bone-like tissues. 

9. Robot Harvests its Own Fuel
A floating, paddling robot (called Row-bot) might be able to run indefinitely with the aid of a microbial fuel cell loaded with bacteria As the robot moves along the surface of water it pulls in a water sample through one entrance, once inside the microbial fuel cell digests or metabolises any organic matter in the water, producing electricity in the process, the water is then expelled, the energy generated is used to move the robot forward and take the next gulp and so repeat every 3 minutes or so. Surplus power might be used for sensors in future. This might also be a water cleanup device in future. But I’ve always been interested in seeing such things developed for medical applications, implanted to power medical devices and harnessing our own blood supply in the process. 

10. Approaches to Anti-Aging Interventions
First, using gene therapy to boost the levels of a membrane-bound cholesterol binding protein on hippocampal neurons was shown to boost neuronal growth and memory in mice, and might show similar benefits in humans Second, overexpression of a single protein was found to allow a population of immune progenitor cells to keep expanding in number when in a cytokine-rich media, and when the cytokines were removed they differentiated into various types of immune cells; they were able to repopulate white blood cells in deficient mice and the process also worked on human cells, a possible important cell therapy in future. Finally, better clarification on the role of GDF11 in parabiosis studies and the potential to restore youthful benefits to aged individuals by boosting the amount in circulation

4 comments on original post
Mark Bruce's profile photoreza heydari's profile photo
The bitter events brought about by blind terrorism in France have once again, moved me to speak to you young people.
In the Name of God, the Beneficent, the Merciful

To the Youth in Western Countries,

The bitter events brought about by blind terrorism in France have once again, moved me to speak to you young people. For me, it is unfortunate that such incidents would have to create the framework for a conversation, however the truth is that if painful matters do not create the grounds for finding solutions and mutual consultation, then the damage caused will be multiplied.
Add a comment...


Shared publicly  - 
The Depth of Commercial Fishing

As technology improves, the places available for fishing increase, but at what cost? +Sam Andrews explores the issues of deep sea fishing and how deep is too deep to be sustainable.

#SciSunCB   #ScienceSunday  
How deep is too deep for commercial fishing?

Our ever-improving technology has allowed us to fish longer, catch more, and move further from land.  It has also allowed us to fish deeper.  EU statistics indicate that between 1950 and 2006 fishing depths increased from an average depth of 407 metres, to 535 metres.

Life in the deep is slow paced.  Food is scarcer than in the sunlit surface waters.  Species grow slower and live longer.  Some deep sea corals, like the one in the image, are thought to be over 4,000 years old. Traits like these are why organisations like +Marine Conservation Institute that ” The deep sea is the world’s worst place to catch fish” .  It’s not just the sustainability of targeted species that is causing concern, but of those caught as bycatch, as well as damage to the seabed and the flora and fauna living in and on it – like the coral in the photo.  So can deep sea fishing ever be managed sustainably?    A recently published study from Joanne Clarke, a PhD student at the +University of Glasgow, and colleagues suggests that there might be a way to make the practice less damaging.

It’s all about depth
The team took a look at trawl survey data undertaken between 1978 and 2013 in the northeast Atlantic.  The surveys used a range of different gear types, were taken at a number of different locations, and varied in depth - between 240 and 1,500 metres.  The data provided the researchers with four ‘indices’ to calculate - Simpson’s diversity index (a measure of diversity), the ratio of “discarded” (bycatch species) to commercial biomass (sellable species), the ratio of Elasmobranchii (sharks and rays) to commercial biomass, and finally the value per square kilometre of each trawl.

Running this data through some fancy analysis, the researchers found some clear patterns.  First, fish diversity increases with depth – an average of 18 new species are encountered for every 100 metres depth you go down.  The ratio of discard biomass to commercial biomass also increased with depth.  At 600 metres the ratio was 0.3:1 but by 1,300 metres, the ratio reached 1.6:1.  The ratio of Elasmobranchii biomass to commercial biomass significantly decreased between depths of 500-600 metres, but then steadily increased again.   The value per unit effort was more variable over depth.  The team found significant decreases in value in trawls between depths of 400–700 metres, but rose again when trawls were around 1,300 metres deep.  The team point to the dominance of rock grenadier (Coryphaenoides rupestris) - a commercially valuable species that lives at these depths as the primary reason for the rise in trawl value at 1,300 metres. 

Based on the results, the researchers recommended that a ban on deep sea trawling would be good for biodiversity, meet a number of conservation targets and EU legislation – including those for deep sea sharks which are believed to be highly vulnerable to deep sea fishing, without much cost to the industry (the obvious exception being any that were to target rock grenadier at depths of 1,300 metres.  If you are wondering if fisheries do go that deep… orange roughy is the most deeply fished species – 1,500 metres below the sea surface).

Politically timely
With deep sea fishing causing widespread concern among academics, NGOs, and those concerned with the environment, fishery managers around the world are being pressed to take action to minimise any impacts this industry can have.  Last year the European Union parliament rejected proposals to completely phase out deep-sea bottom trawling and gill-netting – both considered to be highly damaging activities.  The EU are meeting again this month to discuss how deep sea fishing could be more sustainable – including introducing depth limits to bottom trawling.  The authors hope that the politicians will take note of their findings.   

Are depth limits controversial?
When it comes to fishing, there isn’t much that is not controversial.   Bertie Armstrong, chief executive of the +Scottish Fishermen's Federation responded to this study in a press release.  He is not happy with the idea of banning fishing at depths below 600 metres, and does not see the conservation benefit.  ”Using scientific survey data to draw conclusions upon commercial fishing activity is fraught with danger” Bertie said. ”This is because such information is normally collected via random trawl sampling, whereas commercial fishing tows are very targeted and aimed at specific species or groups of species” .   Bertie argues that commercial boats use much more selective nets that have larger meshes compared to those used in scientific sampling.  He also believes that trawls will become more selective in the future.

The original paper
The paper was published in the journal Current Biology.  The authors have paid for the paper to be open access, so why not have a read of it yourself

The Fishermen’s perspective
Bertie Armstrong’s comments can be viewed on the Scottish Fishermen’s Federation website

The Image 
Like many commercial fishing techniques, bottom trawling produces bycatch of both plants and animals.  Here, a bottom trawler dumps ancient deep sea coral overboard.  Credit © +Greenpeace International/Malcolm Pullman/Marine Photobank

#science #sciencesunday #marinescience #fishing #deepsea #fisheries   #bycatch  
11 comments on original post
Tim Southernwood's profile photoand a's profile photoSam Andrews's profile photo
Thanks for sharing my post +ScienceSunday 
Add a comment...


Shared publicly  - 
Guide To Molting

Molting soon? Perhaps this guide will help you get through it without it ending in disfiguration or death. (For invertebrates only).

#SciSunCB #ScienceSunday  
How to Molt: General Guidelines
Follow this guide and you won't end up like this poor dead sap.

1. Find a nice, quiet spot upon noticing your exoskeleton has become a tad too tight. You will notice an increased rate of protein synthesis in your epidermal cells as hormones are soon released. 

2. You’ll start to feel your epidermis separate from the old endocuticle which is known as apolysis. The name might concern you, but don’t worry – it’s a normal process of molting.

3. Don’t be alarmed by the digestive action of the molting fluid! A special lipoprotein will also be secreted and create the cuticulin layer to protect you before that molting fluid becomes active.

4. The endocuticle of the old exoskeleton will be digested, but your hard work won’t go to waste! Everything will be recycled to form a new procuticle layer underneath your cuticulin layer.

5. Take a deep breath and go! The swelling will cause your old exoskeleton to split open and give you a change of freedom. Be careful, though. It’ll take some time before your new, slightly larger exoskeleton hardens.

6. If you just gained your wings, congratulations! Remember to dry them properly or they'll just be decoration, especially those of you with delicate, folding wings. You'll fly soon enough!

We wish you the best of luck and hope this guide has been helpful for all your molting needs. As long as you time it right and have a perfect chemical balance, you don’t need to fear getting stuck and dying. We’ll see some of you again next time. Until then, stay safe!

Brought to you by #ScienceSunday .
All your science needs, every Sunday.

Sources and Further Reading
Insect Development: Morphogenesis (NCSU | website)
Exoskeletons: The Secret to Insect Success (Cornell | pdf)
12 comments on original post
Hunter Richardson's profile photo
Rule #1:  You are going to be ugly as fuck.  Get used to it.
Add a comment...


Shared publicly  - 
Worth repeating re-sharing
This is an excellent examination of a popular GIF that has been circulating. Thanks +Yonatan Zunger.

#ScienceEveryday   #SciSunCH  
A version of this image was posted [1] with the question: "Will these jellyfish ever make it back to their original place?" +Kimberly Chapman pointed out the obvious "yes, because it's an animated GIF and those loop." But here's something you might not expect: even if it weren't an animated GIF, even if these jellyfish were being moved around by a program using random numbers, I could guarantee that they always repeat.

(Or to be a bit more careful, if you watch it long enough, you'll always see a repetition. It might not be the very first position that repeats)

Why? Let's imagine a simpler case for a moment, involving a 2x2 grid of jellyfish, each of a different color so we can tell them apart. There are 24 possible ways we could arrange the jellyfish: if you start with an empty grid, there are four places to put the red jellyfish; for each of those places, there are three remaining possible places to put the blue one; for each of those, there are two remaining possible places to put the green one; and once you've chosen those three, there's only one place the yellow one could go. So there are 4*3*2*1=24 possible jellyfish patterns. (Written 4!, "4 factorial")

Each time the jellyfish move, we move from one of these 24 configurations into another. As it happens, the motions below are very limited -- each jellyfish has to move onto a dot next door -- but that turns out not to matter, because even if the jellyfish could teleport, they'd still have to repeat.

Why? Imagine that we look at the first 25 moves. The jellyfish will end up in 25 configurations, but there are only 24 different configurations total, which means that at least one configuration had to happen twice!

This is called the pigeonhole principle: if you have N+1 pigeons in N pigeonholes, at least one hole has to contain two pigeons. (In this case, you have 25 configurations in 24 distinct slots)

If the rule going from one configuration to the next is deterministic -- that is, if the next move depends only on where the jellyfish are right now -- then you know that once a single repetitive loop happens, that loop will continue to repeat forever, because you're back at the first stage of the loop and will then have to go on to the second one, etc.

If the rule isn't deterministic -- say, if each time the jellyfish move randomly -- then no particular pattern is guaranteed to repeat, but you know that no more than once every 25 moves, at least once the jellyfish will repeat their pattern.

The same thing is true for this bigger grid; you just need to wait a bit longer. The 16x16 grid below has 256 jellyfish, so you need to wait for 256!+1 steps -- that's 256*255*...*3*2*1 + 1 steps, or about 8*10^506 steps [2] -- but no matter what, the jellyfish are absolutely guaranteed to repeat.

What's even more interesting is that this may apply to more than just jellyfish. One set of rules that we know are deterministic are the laws of physics. [3] Now, an interesting open question in physics is: is there a minimum granularity of spacetime, so that we can think of the entire universe as being on some kind of extremely fine grid? (When I say "extremely fine," I mean a grid size of the Planck length, about 1.6*10^-35 meters. For comparison, that's as much smaller than a proton as a proton is smaller than the San Francisco Bay Area.)

There are some reasons to believe that this may actually be true (although the geometry is a lot more complicated than a simple grid, and in fact "geometry" isn't even the right word for it; the whole expansion of the universe, from the big bang on, is part of it). If it is, then there's something interesting: we could imagine the entire universe as a gigantic grid, and the current state of the universe is given by deterministic laws about what's on that grid, then we know that the state of the universe itself must ultimately repeat.

Of course, "ultimately" is a pretty long time horizon: if you think the number below is big, that's what we got with only 256 jellyfish. The total number of "jellyfish" needed to describe the universe is going to be something like 10^245, and so the number of moves it would take would be unimaginably huge.

But if this repetition is real, then it has some very interesting consequences. For example, it's one way to explain why we happen to observe physical constants in our universe that are consistent with the existence of human life. [4] If those "constants" are actually controlled by the state of the universe, and the universe ultimately steps through all possible states, then it isn't surprising that we'll look out the window and see the constants that we could survive in; when the universe was in all of those other possible states, we weren't around to see it.

If, on the other hand, the universe has infinitely many states in it, then no recurrence need ever happen; it can keep changing indefinitely, and the entire argument above falls apart. This is one of the very few times that "finite but very big" and "infinite" are meaningfully different in physics.

This sort of analysis is called an "anthropic" analysis, and while it seems unsatisfying in some ways -- it doesn't explain the values of the constants, after all, or tell us what other constants might allow us to exist, it just tells us why they happen to be that right now -- it's a real possibility that this is what's actually going on. The entire debate over this, whether these recurrences (they're called Poincaré Recurrences, after the French mathematician who first described the math above) occur in nature and whether Anthropy is an explanation for the world, is a major open question in fundamental physics today.

So whether the jellyfish are moving in an animated GIF or powering the basic laws of physics, remember this: Finite patterns must always repeat; infinite patterns don't have to.

And now, you may return to staring at the GIF to your heart's content.

[1] By +AsapSCIENCE, but without real source credit. I strongly suspect, but cannot verify, that this is the work of the great Dave Whyte of, which if you like GIFs like these you should absolutely check out. If anyone knows where this image actually came from, I'd love to know.

[2] If you want to be precise about it, it's 857,817,775,342,842,654,119,082,271,681,232,625,157,781,520,279,485,619,859,655,650,377,269,452,553,147,589,377,440,291,360,451,408,450,375,885,342,336,584,306,157,196,834,693,696,475,322,289,288,497,426,025,679,637,332,563,368,786,442,675,207,626,794,560,187,968,867,971,521,143,307,702,077,526,646,451,464,709,187,326,100,832,876,325,702,818,980,773,671,781,454,170,250,523,018,608,495,319,068,138,257,481,070,252,817,559,459,476,987,034,665,712,738,139,286,205,234,756,808,218,860,701,203,611,083,152,093,501,947,437,109,101,726,968,262,861,606,263,662,435,022,840,944,191,408,424,615,936,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,001 steps.

[3] You may have heard something about "quantum randomness," but this isn't actually randomness; the actual evolution of wave functions is completely, 100%, deterministic, even in quantum mechanics.

[4] The Standard Model of particle physics is controlled by about 20 basic constants, like the mass of the electron and the strength of gravity. Our world is weirdly sensitive to some of them: if the down quark were 10% heavier, say, then stars would never form, and neither would nearly any other kind of matter. What controls these 20 values? Good question. We don't know yet.
131 comments on original post
Add a comment...
Have them in circles
87,340 people
Alexander Rauscha's profile photo
manish hirapara's profile photo
Odair Alfredo Gomes's profile photo
Yasmin Pukin's profile photo
Finn Waage Michaelsen's profile photo
Dawn Mason's profile photo
Sahda Haroon's profile photo
rebecca kent's profile photo
Dan Epstein's profile photo


Shared publicly  - 
SciTech digest 52
Thanks as always +Mark Bruce 

SciTech #ScienceSunday Digest - 52/2015.
Permalink here:
Merry Xmas and Happy New Year to all from SciTech Digest! 

SpaceX rocket success, Seabed mining robots, Cheaper DNA synthesis, Machine learning projects, Boosting brain cleaning, Electron beam nanofabrication, Optoelectronic chips, Intelligence gene networks, Plant virus vs cancer, Protein self assembly. 

1. SpaceX Successfully Lands Reusable Falcon 9 Rocket
In easily the biggest news of the week SpaceX successfully landed its reusable Falcon 9 rocket after launching 11 communications satellites into orbit After two prior landing failures and a launch failure this milestone is a welcome, if not entirely unexpected, demonstration. It marks the first time an orbital rocket has delivered a payload into orbit and returned to Earth unscathed and ready for refurbishment and relaunch. Further tests to replicate the feat and build on the capability will result in drastic reductions in the cost of ferrying mass to space and help usher in new space-based applications and capabilities for our species. 

2. Robots that Mine the Seabed
The first deep sea seabed mining robots have been developed and will be tested throughout 2016-17 until the arrival of their first deployment and processing ship, which is currently under construction After ripping up ore from the seafloor dredge pumps carry the rock to the surface where the deployment ship removes the water and transfers the ore to another ship for transport to China for processing (I do wonder if this step can ever be co-located with the main ocean operations). Over 30 months the robots will bring up 2.5 million tons of ore with metals worth $1.5 billion. 

3. DNA Synthesis Races to the Bottom
New start-up companies are developing technology to manufacture and synthesise DNA sequences faster and cheaper than ever A key part of this trend is increasing automation of the synthesis process with one example being an automated synthesis process taking part in each of the 10,000 tiny etched 600-nm wells adoring a new silicon synthesis chip that is fabricated with standard photolithography processes and will start providing customers with DNA at 10 cents per base pair but aiming for 2 cents in the foreseeable future. While cheap synthesis doesn’t help quite so much once you have a desired cell and application, cheap synthesis does help greatly with rapid iteration, accelerated experimentation, and distributed development. 

4. Big Week for Machine Learning Applications
First, a new machine learning system based on new object-recognition algorithms that more closely resemble how the human brain recognises objects requires only a few exposures rather than millions Second, computers are being taught via human crowdteaching how to successfully tackle difficult driving tasks, and this might open up similar avenues to teach robots etc how to master other non-trivial tasks Third, SegNet is a new system able to run on a normal camera or smartphone able to efficiently and quickly identify location, orientation, and various components in the scene ahead to better empower autonomous vehicles Finally, a new deep-learning algorithm can predict an images memorability at near-human levels

5. Boosting Brain Cleaning System Boosts Cognition
Rolipram, a drug originally developed in the 1990s but which never made it to market due to nausea-inducing side effects, has been found to rescue and activate the proteasomes in neurons and helps to boost cognition in mouse models The proteasome is an enzyme that degrades old proteins for recycling and its inactivity is involved in neurodegenerative diseases characterised by a buildup of old toxic proteins; better, safer drugs and other interventions building on this work will be attractive options for treating Alzheimers, Parkinsons, and Huntingtons Disease, as well as helping to generally rejuvenate cells and tissues throughout the body. 

6. Electron Beam 3D Nanofabrication
A liquid rather than gaseous precursor has helped speed up electron beam 3D nanofabrication by up to 5,000 times, finally helping to scale this promising technology The proof of concept demonstrated micrometer-tall nanopillars and suspended 3D nanostructures; the new technique and new precursors should allow for the rapid development and fabrication of electrode surfaces for batteries and fuel cells, vertically-stacked electronic memory, novel substrates for cell manipulation, and nano-micro-scale electrochemical conversion devices. 

7. Latest Optoelectronic Computer Chips
Optical interconnects in integrated circuits have been a long time coming but this latest effort sees genuine commercial scale integrated circuits built using existing microchip fabrication facilities and processes that incorporate new photonic / optoelectronic interconnects; the transistors compute information conventionally but the chip uses light to move information around the chip much more efficiently The demonstration chip has 70 million transistors and 850 optical components and is capable of the functionality required of a typical commercial chip, and with a relatively clear path to get to the billion transistor scale needed for modern chips. Lots of interesting materials and chip design innovation here. 

8. Gene Networks Linked to Human Intelligence
A Big Data analysis of gene expression profiles from human brain samples, genomic information from both healthy people who had completed IQ tests and also those with neurological disorders has identified two gene networks, each with hundreds of genes, that significantly influence intelligence and cognitive function in the human brain Both networks appear to be under the control of master regulator switches that the group are now seeking to identify. The group hopes the knowledge gained will help in the development of treatments for neurological disorders and also powerful enhancers for healthy cognitive function. 

9. Plant Virus Clears Mammalian Cancer Cells
Turns out that the shells of common plant viruses (that have had their DNA removed) induce significant immune responses against cancer cells in mice This demonstration was successful in lung tumours when the viral shells were inhaled and also when injected into ovarian, colon, or breast tumours, in all cases igniting immune responses to clear the tumours and protect the animal against systemic metastases. Such in-situ vaccinations are a 100 year old idea and in this case proved to be a very effective switch in activating the immune system to recognise and destroy the tumour cells - essentially a biological adjuvant. 

10. Structural Nanotechnology: Protein Sheets & Shells
Bacterial proteins have been observed self-assembling into sheets, and curved sides of microcompartments as they form for the first time The group observed hexagonal protein subunits with a concave or convex structure joining together to form larger curved structures and modelled the dynamics of this behaviour. It is known that bacteria use such self-assembled protein microcompartments to isolate certain reactions and enzymes; such knowledge might help in the design of artificial microcompartments to perform custom reactions inside cells and other structures. 

14 comments on original post
Mark Bruce's profile photo
Cheers Chad and all the best for the new year! 
Add a comment...


Shared publicly  - 
Have a heart
#ScienceSunday   co-curator +Chad Haney shares and interesting news byte about medical imaging and hearts from an excavation.

Heart of a Knight
I meant to post this back on 3 December when the story was fresh. I was hoping more information would become available. However, it seems that it was only reported at the recent RSNA meeting and that's it. I'm very disappointed that they didn't show any CT or MRI images.

A group of imaging experts were called when five heart shaped urns were found during an excavation of the basement of the Convent of the Jacobins in Rennes, France. Several grave sites dating back to the late 16th or early 17th century were unearthed. The initial CT and MRI images were good but not diagnostic. The embalming fluid likely has alcohol which would cause chemical shift artifacts (read my MRI post to learn more). The CT would not have good soft tissue contrast and therefore would not have been helpful by itself. After removing the embalming fluid and "re-hydrating" the hearts, one was found to be healthy, three had signs of plaque on the coronary arteries and atherosclerosis and the fifth was poorly preserved.

I was also considering saving this until St. Valentine's day as they mentioned that one of the male hearts was buried with a female, likely his wife. You can read the whole news blurb here:  

Medical Imaging 101 pt 2: CT

Medical Imaging 101 pt 3: MRI

17 comments on original post
Add a comment...


Shared publicly  - 
Red meat, what's the real risk?
ScienceSunday Co-curator +Buddhini Samarasinghe breaks down a bit of the hyperbole swirling around the classification of red meat and processed meat as cancer causing. You're probably thinking this is old news, 'I remember something about charred meat'. This story is slightly different so read on.

However, here's what +Siromi Samarasinghe had to say about cooking meat.
Method of cooking meat has also been linked to cancer risk. Irrespective of the type of meat, exposure to high temperatures favour the formation of certain chemicals known as Polycyclic Aromatic Hydrocarbons (PAH) and Heterocyclic Amines (HCA) which can cause changes in DNA. Grilling, barbecuing, exposure to smoke and charring may increase the formation of these compounds.   Moderation is best!
#ScienceEveryday   #SciSunCH  
Red meat and cancer risk

The news is awash with stories about how red and processed meats have been classified as carcinogens in the same category as tobacco. But what exactly does this mean? Let's unpick this a little bit before throwing out the bacon with the bathwater. 

There have been several excellent bits of writing that explain what this means - the first is by Ed Yong ( and the second by CRUK* ( These are well-worth a read if you want to learn more. 

Basically, the key bit of information to remember is that this is not a risk assessment, it is a hazard identification. A great analogy (stolen from the CRUK article above) is to think of banana skins - they definitely can cause accidents, but in practice it doesn't happen very often, and isn't as severe as being in a car accident. But under the hazard identification approach, banana skins and cars would be in the same category because they both definitely cause accidents. The severity of the accident is not discussed, and that's where we tend to get lost with the breathless press releases on this topic. 

So should you stop eating red and processed meat? The answer is all about the dreaded, boring M word - moderation. If you're always eating red and processed meat, over years and years, then that's probably not good for you. But meat in moderation (i.e. not too much and not too often) is still okay, and is definitely not as bad as smoking is. The thing with diet and disease is that reality is often rather boring; there are no miracle diets or magical juice cleansers that will give you eternal youth. There are no superfoods that offset the damage of binge-drinking every weekend. That's just not how our bodies work. 

What you can do to prevent cancer is eat plenty of fruit and veg with lots of fibre while cutting back on things like alcohol, salt, red and processed meats. And definitely avoid sunburns and smoking. 

*In the interest of full disclosure, I work at the charity CRUK as a science communicator. 

52 comments on original post
Robert Clark's profile photo
How many lethargic people eat hotdogs. Could laziness be the cause of cancer?
Add a comment...


Shared publicly  - 
SciTech Digest

Another weekly science digest courtesy of +Mark Bruce.

#SciSunCB   #ScienceSunday  
SciTech #ScienceSunday Digest - 36/2015.
Permalink here: 

Bacterial produced antibodies, Physics beyond QM, Graphene wonders, Biological computer circuits, De novo genes, Autonomous pest killers, Neuromemristive processor, Miniature LIDAR, Metasurface optics, Self-assembled nanowires. 

1. Bacterial Production of Human Antibodies
Monoclonal antibodies for therapeutic and research use are typically manufactured in sensitive and expensive animal cell lines, but with newly engineered E. coli bacteria monoclonal antibodies and other complex therapeutic proteins can be produced quickly, cheaply, and in therapeutic dosages The bacteria contain the genetic instructions and cellular machinery needed for human antibody production and proof-of-concept experiments produced full-length antibodies that were effective against flu, anthrax, and the breast cancer drug Herceptin. Modifications to produce different proteins as needed can be quickly made. This seems like a very powerful, distributed drug production platform. 

2. Physics Beyond the Standard Model
A recent analysis of data from the Large Hadron Collider suggests that the heavier electron-lepton particles (muon and tau) are behaving in a manner that the Standard Model cannot fully account for The key measure concerned differences in decay rates and suggests lepton universality in the Standard Model no longer holds, and was further replicated in two different experiments. Worth taking with a grain of salt of course, but will be interesting to see if future, more powerful experiments confirm the result. 

3. Interesting Graphene and Nanotube Advances
Graphene has been made to superconduct after being doped with lithium, with electrons in the material actually slowing down compared to standard room temperature graphene, but by doing so seem to form the classic Cooper-pairs that typically characterise superconductivity In related carbon news (i) nanotubes measuring just 1nm wide have been used as templates to grow long one-dimensional diamond polymers within their tubes that can be subsequently extracted and recovered, (ii) a new layering process forms 3D graphene structures, and (iii) newly discovered phagraphene exhibits interesting directional ballistic electron properties

4. Biological Computer Circuits
Two strains of genetically engineered bacteria have been created that jointly regulate the production of proteins needed for intercellular communication The strains cooperated to up-regulate and down-regulate genes of interest, creating positive and negative feedback loops that affected gene expression across the entire population of bacterial cells. The ultimate applications for this work will involve engineered cells that take up residence in your gut flora for example and can be remotely triggered via dietary additives to modify the regulatory behaviour of a larger population of different strains in your gut. 

5. The Phenomenon of De Novo Orphan Genes
This is a good overview article about de novo orphan genes, genes that randomly appear with no resemblance to any other gene and no known relatives, seeming to appear from chance changes to junk or non-coding DNA Once considered to be impossible, such genes are being found far more frequently and there are now 600 human candidate genes. Think about that: random sequences of “junk” DNA accumulate mutations that (i) allows it to be read as a gene, (ii) includes upstream regulatory sequences for appropriate activation, (iii) can translated into RNA and protein, and (iv) that protein has a useful function (plugs into existing networks) in order to ensure the new gene is retained. 

6. Autonomous Robot Sea Star Killers
COTSBot is an autonomous underwater robot, shaped like a 30kg yellow torpedo, capable of swimming at 2 m/s for 6 hours, with 5 independent thrusters that allow stable hovering in a single spot, with an automatic machine-vision enabled crown of thorns sea star image and object recognition system, and an automatic poison dart injection device for rapidly killing sea stars The device is intended to help clear coral reefs of damaging crown of thorns sea star infestations autonomously and to keep sea star populations down to improve reef health; the poison was developed over many years to be specific to sea stars and doesn’t harm other reef inhabitants. I wonder how long until similar systems are adapted to other types of autonomous drones to target other invasive land pests - cane toads in Australia for example. 

7. Knowm’s Neuromemristive Processor
The company Knowm has developed an adaptive neuromemristive neuromorphic computer processor intended for applications in machine learning and autonomous platforms The architecture eliminates the separation of memory and processing, utilises Knowm’s new “kT-RAM” processor and synapses, and are now being made available to people to test and prototype along with Knowm’s API and emulators. An interesting entry into the space along with giants such as IBM and others. 

8. Miniature LIDAR Systems
A new, fast, self-sweeping laser system can dramatically reduce the power consumption, size, weight, and cost of LIDAR and optical coherence tomography devices compared to those available today - even to the point of being mounted on small drones and smartphones. The key innovation is the integration or coupling of the laser with an ultra thin mirror that harnesses the force of light to move the mirror, and allowing several million laser sweeps per second to enable 3D image capture for realtime videos. In related news people can now spoof current LIDAR sensors on autonomous cars for example to make them think certain objects are present at certain points

9. Metasurface Optics and Printable Holograms
A metasurface made out of tiny pillars of silicon acts as a waveguide for light, allowing the control of focus and polarisation and comprising a new platform for optics Such a metasurface might replace conventional optics in future with with far less bulk to allow thinner, lighter phone camera lenses, better control over industrial laser systems and possible novel applications that include holographic and 3D displays viewable without glasses. In related news holographic lenses can be printed for a range of applications

10. Custom Nanowires from DNA and Protein
Nanowires made out of specific small proteins and DNA subunits that self-assemble in solution are a nice example of a material that was first designed on a computer One of the reasons I like this advance is its cusomisability; you can stick with assembling one-dimensional polymer strands to arbitrary lengths, but by simply extending the DNA sequence you form custom crosslinks between strands and build up a network of nanowires of arbitrary density, while adding other peptide subunits onto the proteins would allow further crosslinking and structural tweaks to self-assemble a novel biocompatible mesh material. 

11 comments on original post
Add a comment...


Shared publicly  - 
How did veggie munching caterpillars evolve into carnivores? +Tommy Leung tells us about the caterpillars of the blue butterfly. 
Alice Ant And The Curious Carnivorous Caterpillar
This is a piece which I drew for a radio segment here: where I talked about carnivorous caterpillar.
There are over 160000 species of known moths and butterflies and most of them are herbivorous. In fact, caterpillars are very specific about what plants they munch on - some only ever feed on a single species of plant throughout their development, and few have a host plant range of more than three plant families.
But there is a tiny percentage of caterpillars that have evolved to be carnivorous, about one percent of all know lepidopteran (moths and butterflies) are actually carnivores. They have all evolved to be that way through different paths - some of them from associating with ants and their aphid livestock, others through becoming isolated and evolved into a predatory niche in the absence of other carnivorous insects. To find out more about the fascinating story behind carnivorous caterpillars, you can listen to the radio segment I recorded yesterday on ABC New England for their segment "Creepy but Curious" here:

Earlier today, I recorded yet another episode of ;Creepy but Curious ;- a monthly radio segment that I have been doing on the local radio since 2013. And with every segment that I do, I als...
2 comments on original post
Add a comment...


Shared publicly  - 
Add a comment...


Shared publicly  - 
Being 50 and pregnant isn't a big deal for sharks

50 and pregnant? No big deal—if you're a great white shark.

A recent study found that great white females mature later than once thought—in their mid-30's—and once they reach sexual maturity, great whites can reproduce throughout their lives. Researchers also found that great whites live longer than expected, at least 73 years. 

A recently-sighted 20-foot great white named Deep Blue, thought to be the biggest white shark ever filmed, was estimated by shark experts to be around 50 years old, based on her size. Her unusual girth suggested that she was also pregnant. Larger, older sharks typically produce bigger litters, making sizable females like Deep Blue especially important for maintaining shark populations in the wild.

Happy #ScienceSunday!
A marine scientist is dwarfed by what is being called the biggest shark ever caught on camera in a recent video taken near Mexico's Guadalupe Island. How did this 20-foot-long (6 meters) great white Internet sensation become such a behemoth?
View original post
Add a comment...