Recently Browsing 0 members
No registered users viewing this page.
By Ather Fawaz
Researchers probe into RNA using deep learning to develop sensors for a COVID-19 diagnostic
by Ather Fawaz
A genome is a genetic blueprint that determines an organism's characteristics. Deoxyribonucleic acid (DNA), and usually in the case of viruses, Ribonucleic acid (RNA) are the building blocks of genomic sequences. And manipulating these nucleic acids directly can lead to tangible changes in the organism.
As such, developments in genetic engineering focus on our ability to manipulate genomic sequences. But this is a daunting task. For example, precisely controlling a specific class of engineered RNA molecules called "toehold switches" can lend vital insight into cellular environments and potential diseases. However, previous experiments have shown that toehold switches are not tractable, many don't respond to modifications even though they have been engineered to produce the desired output in response to a given input based on known RNA folding rules.
Considering this, two teams of researchers from the Wyss Institute at Harvard University and MIT have developed a set of machine learning algorithms that can improve this process. Specifically, they used deep learning to analyze a large volume of toehold switch sequences to accurately predict which toeholds perform their intended tasks reliably thereby allowing researchers to identify high-quality toeholds for their experiments. Their findings have been published in Nature in two separate papers today.
With any machine learning problem, the first step is to collect domain-specific data to train the model on. The researchers collected a large dataset composed of toehold switch sequences. Alex Garruss, co-first author and a graduate student working at the Wyss stated:
Since there were two separate teams, the researchers tried their hands with two different techniques to approach the problem. The authors of the first paper decided to analyze toehold switches not as sequences of bases, but as 2D images of base-pair possibilities. This approach, called Visualizing Secondary Structure Saliency Maps, or VIS4Map, successfully identified physical elements of the toehold switches that influenced their performance, providing insight into RNA folding mechanisms that had not been discovered using traditional analysis techniques.
After generating a data set of thousands of toehold switches, one team used a computer vision-based algorithm to analyze the toehold sequences as two-dimensional images, while the other team used natural language processing to interpret the sequences as "words" written in the "language" of RNA. Image via Wyss Institute at Harvard University Authors of the second paper created two different deep learning architectures that approached the challenge of identifying 'susceptible' toehold switches using orthogonal techniques. The first model was based on convolutional neural network (CNN) and multi-layer perceptron (MLP), that treated the toehold sequences as 1D images, or lines of nucleotide bases. Using an optimization technique called Sequence-based Toehold Optimization and Redesign Model (STORM), it identified patterns of bases and potential interactions between those bases to mark the toeholds of interest.
The second architecture modeled the problem to the domain of natural language processing (NLP), treating each toehold sequence as a phrase consisting of patterns of words. The task was then to train a model to combine these words, or nucleotide bases, to make a coherent phrase. This model was integrated with the CNN-based model to create Nucleic Acid Speech (NuSpeak). This optimization technique redesigned the last nine nucleotides of a given toehold switch while keeping the remaining 21 nucleotides intact. This allowed for the creation of specialized toeholds that detect the presence of specific pathogenic RNA sequences and could be used to develop new diagnostic tests.
By using both models sequentially, the researchers were able to predict which toehold sequences would produce high-quality sensors. Image via Wyss Institute at Harvard University To test both models, the researchers sensed fragments from SARS-CoV-2, the viral genome that causes COVID-19, using their optimized toehold switches. NuSpeak improved the sensors' performance by an average of 160%. On the other hand, STORM created better versions of four SARS-CoV-2 viral RNA sensors, improving their performance by up to 28 times. Apropos these impressive results, co-first author of the second paper, Katie Collins an MIT student at the Wyss Institute, stated:
Diogo Camacho, a corresponding author of the second paper and a Senior Bioinformatics Scientist and co-lead of the Predictive BioAnalytics Initiative at the Wyss Institute stated:
Moving forward, as Camacho envisioned, the teams are looking to generalize their algorithms to map them onto other problems in synthetic biology to potentially accelerate the development of biotechnology tools.
Folding at Home now the fastest "computer" in the world, but also join our team
by Christopher White
It's obvious that there's nothing good about the coronavirus itself. However one positive has been the outpouring of support for the Folding@Home project that's looking for a cure to diseases such as Alzheimer's, cancer, and COVID-19. Last week we asked our readers to join the project (and the Neowin team, 55186), and you responded. We now have over 200 new folders on the team and in the last week, have jumped up 90 spots to rank 661 overall.
Neowin readers aren't the only ones responding to this crisis. As noted on Tom's Hardware, the F@H project has more compute power than not just the fastest supercomputer in the world, but the top seven supercomputers in the world, combined. Since the coronavirus outbreak, the project has seen a 1,200% increase in the number of folders, with over 400,000 people joining. The total number of CPU/GPU cores being used by the project is 27,433,824.
A work unit crunching away at the Coronavirus problem We'd love if you joined the Neowin team. Simply install the client, type in a username, enter team number 55186, and you'll be folding with us in no time! If you're not seeing many work units assigned to you right now, keep in mind that due to the outpouring of support, the scientists need to provide more data for our computers to crunch so it's possible your PC will idle for a bit until they get this worked out. Just leave the application running, and when work comes in, you'll be folding in no time!
We also have a dedicated forum thread discussing the project right here.
Alphabet's Verily establishing Covid-19 testing sites in California
by Paul Hill
Image via Wikimedia Alphabet’s Verily has announced that it is working with the California Governor’s office, federal, state, and local public health authorities to create testing sites in the Bay Area for those at high risk of contracting Covid-19. It is also building an online tool which will help health workers figure out who is infected and needs treatment.
The online test, which is part of Project Baseline, is now available to Californians that are interested in getting tested for the illness. Right now, the availability is very limited with testing being offered to Californians located in Santa Clara County and San Mateo County. The firm has said that it’s working quickly to expand testing in every way that it can including with added test sites and expanded eligibility criteria.
Discussing the reason for beginning at the sites it has, Verily said:
As time progresses, Verily hopes to expand its testing access across more areas in California. It said the program is still in the early stages and that it’ll take time to asses operations at pilot sites before it rolls out elsewhere. If you’re in California, Project Baseline recommends that you follow advice from the CDC website.
If you have spare computing power and want to help fight the Coronavirus, think about joining the Neowin Folding@Home team and allow your PC to crunch data.
Join the Folding at Home Neowin team to fight the novel Coronavirus
by Christopher White
The Folding at Home project has been around for two decades and is still going strong. For the uninitiated, the project conducts disease research by carving out units of work that can be shipped to an individual's computer so that those machines can conduct protein folding simulations. When your computer is done crunching the numbers, it sends the results back to the Folding at Home servers and requests another unit of work.
The group has recently started assisting scientists in finding a cure for the novel Coronavirus, COVID-19. What this means is that your spare CPU cycles can be donated to the project to help find a cure to the pandemic that's impacting everyone's lives around the world. The project is aiming to recruit a million volunteers.
Helping out is easy: Simply download the program from their website, type in what name you want to use and optionally what team you want to join, and let it go. You can configure how much machine power you want to donate, and you can even click on the Configure button to setup how many CPU cores you want to provide. As a warning, if you let it consume your entire machine, it will definitely peg the CPU at 100% and generate quite a bit of heat. My workstation is powered by the Ryzen 3900x, and after initially giving the tool access to all 24 cores, I noticed the CPU temperature was extremely hot, so I limited it to only 12 of the cores, which is still plenty. The tool can also use your PC's GPU for even more processing, and that's currently the method used for the COVID-19 tests. You can search any of the projects to find out who is using the research and what it's for on the Folding at Home website.
Neowin has had our own team since 2007, so when doing the install, it'd be great if you used our team number: 55186. The front-end servers are getting hammered recently with thousands of people rushing to sign up and help fight the disease, so you'll often receive a "Bad Gateway" error when checking, but when things are working, you can check the status directly on the Folding at Home page by typing your name or team number into the search box.
Google asks all North American employees to work from home
by Rajesh Pandey
Google has issued a new notice to its employees in North America, recommending them to work from their home through April 10. Google's vice president of global security Chris Rackow has sent out an email to all of the company's employees about this. Google has 11 offices in the United States and Canada that are affected by this directive.
The notice comes just a week after the company had told its employees based in Washington to work remotely. Apart from America, Google has issued similar guidance for its employees in different parts of the world including Europe. If their positions allow for it, all Google employees are told to work remotely.
Below is what Rackow's email to Google employees read:
Google is also carefully monitoring the entire situation and will provide an update with a relevant timeline as required.
Additionally, Google has also formed a COVID-19 fund that will enable its temporary staff and vendors globally to take paid sick leave if they show symptoms of the virus or are quarantined. This fund will allow Google to compensate its workers for normal working hours if they are unable to come to work for the above reasons.