mjpdejong
Shared posts
DNA in blood can track cancer development, response in real time
Facebook werkt aan deep-learningsysteem dat go-spelers moet verslaan
Direct Genomics' New Clinical Sequencer Revives a Forgotten DNA Technology
Online Ancient Genome Repository (OAGR)
Vitamin D pill a day may improve exercise performance and lower risk of heart disease
Een einde aan strikte netneutraliteit: trager YouTube en sneller Netflix?
Silicon Valley's thrust into biotech sparks scramble for data scientists
The search at Apple, Google and other members of the tech elite for new business opportunities is pushing them to recruit from outside of their core fields. And whether the sector being targeted is artificial intelligence, self-driving cars or life sciences, the result is the same: a tug-of-war with incumbents for talent.
The science behind real life zombies
Lafferty KD. (2006) Can the common brain parasite, Toxoplasma gondii, influence human culture?. Proceedings. Biological sciences / The Royal Society, 273(1602), 2749-55. PMID: 17015323
Can the common brain parasite, Toxoplasma gondii, influence human culture?Vyas A, Kim SK, Giacomini N, Boothroyd JC, & Sapolsky RM. (2007) Behavioral changes induced by Toxoplasma infection of rodents are highly specific to aversion of cat odors. Proceedings of the National Academy of Sciences of the United States of America, 104(15), 6442-7. PMID: 17404235
Behavioral changes induced by Toxoplasma infection of rodents are highly specific to aversion of cat odors.Passamonti L, Crockett MJ, Apergis-Schoute AM, Clark L, Rowe JB, Calder AJ, & Robbins TW. (2012) Effects of acute tryptophan depletion on prefrontal-amygdala connectivity while viewing facial signals of aggression. Biological psychiatry, 71(1), 36-43. PMID: 21920502
Effects of acute tryptophan depletion on prefrontal-amygdala connectivity while viewing facial signals of aggression.Thomas, F., Schmidt-Rhaesa, A., Martin, G., Manu, C., Durand, P., & Renaud, F. (2002) Do hairworms (Nematomorpha) manipulate the water seeking behaviour of their terrestrial hosts?. Journal of Evolutionary Biology, 15(3), 356-361. DOI: 10.1046/j.1420-9101.2002.00410.x
Do hairworms (Nematomorpha) manipulate the water seeking behaviour of their terrestrial hosts?W. Wesołowska T. Wesołowski. (2014) Do Leucochloridium sporocysts manipulate the behaviour of their snail hosts?. Journal of Zoology , 292(3), 151-155. info:/10.1111/jzo.12094
DevOps case study: Rabobank cuts deployment by 60 percent
CEO of XebiaLabs shares a DevOps success story from the front lines.
Gene therapy treats all muscles in the body in muscular dystrophy dogs
Yue, Y., Pan, X., Hakim, C., Kodippili, K., Zhang, K., Shin, J., Yang, H., McDonald, T., & Duan, D. (2015) Safe and bodywide muscle transduction in young adult Duchenne muscular dystrophy dogs with adeno-associated virus. Human Molecular Genetics, 24(20), 5880-5890. DOI: 10.1093/hmg/ddv310
Safe and bodywide muscle transduction in young adult Duchenne muscular dystrophy dogs with adeno-associated virusSELEX: How It Happened and Where It will Go.
Related Articles |
SELEX: How It Happened and Where It will Go.
J Mol Evol. 2015 Oct 20;
Authors: Gold L
PMID: 26480964 [PubMed - as supplied by publisher]
Dead men punching
Venter's Human Longevity starts $50,000 health testing service
J. Craig Venter has unveiled the scope of the Health Nucleus testing service being offered by Human Longevity Inc. For $50,000, HLI will perform a battery of sequencing, imaging and laboratory tests to build a picture of the health of an individual--while adding a wealth of information to its burgeoning database.
Quantum physics meets genetic engineering
Bandage: interactive visualization of de novo genome assemblies
Summary: Although de novo assembly graphs contain assembled contigs (nodes), the connections between those contigs (edges) are difficult for users to access. Bandage (a Bioinformatics Application for Navigating De novo Assembly Graphs Easily) is a tool for visualizing assembly graphs with connections. Users can zoom in to specific areas of the graph and interact with it by moving nodes, adding labels, changing colors and extracting sequences. BLAST searches can be performed within the Bandage graphical user interface and the hits are displayed as highlights in the graph. By displaying connections between contigs, Bandage presents new possibilities for analyzing de novo assemblies that are not possible through investigation of contigs alone.
Availability and implementation: Source code and binaries are freely available at https://github.com/rrwick/Bandage. Bandage is implemented in C++ and supported on Linux, OS X and Windows. A full feature list and screenshots are available at http://rrwick.github.io/Bandage.
Contact: rrwick@gmail.com
Supplementary information: Supplementary data are available at Bioinformatics online.
Mini DNA sequencer tests true
Sequencing the Genomes of Dead People
Last Wednesday, at “Career Night” during the American Society of Human Genetics annual conference in Baltimore, I was stationed one table over from Robert Steiner, MD, from the Marshfield Clinic Research Foundation in Wisconsin. With young scientists circling us like electrons around nuclei, I never got the chance to break away to talk to him. But I did overhear him discussing the Genomic Postmortem Research Project, an effort to sequence the genomes of 300 dead people.
I was fascinated.
Would knowing the information encoded in the DNA of the deceased have changed their health care? I went to the talk on the project the next day to find out about this clever test of the value of genome sequencing.
A BRIEF HISTORY OF HUMAN GENOME SEQUENCING
Genomes transcend time. The beauty of seeking clues to health in DNA sequences is that this relatively immutable record is present in nearly every cell of an individual. Two copies of our genomes are nestled into the nuclei of cells of our kidneys and spleens, bone marrow and eyeballs, the same blueprint stamped over and over down the cell lineages at every mitosis, with a few somatic mutations spewed here and there. The genome in the fertilized ovum closely matches that in the cells of the 90-year-old it might someday become, given good genes and a lot of luck. And DNA rests within the bone cells of the skeleton long after death, the mitochondrial satellite source even more resistant to time and trauma.
“The” human genome was first sequenced in 2001 or 2003, depending upon which draft is cited, and included a compendium of DNA donors from the International Consortium and the lone genome of J. Craig Venter from the Celera Genomics version. The genome sequence of another rock star of science, James Watson, came next. Then came a few “firsts,” such as the genome of a man called simply “YH,” a Han Chinese. Representatives of other population groups followed.
Then came the famous. The late Steve Jobs and Christopher Hitchens were sequenced in futile attempts to combat their cancers. Henry Louis Gates Jr. did it to trace his African roots. The actress Glenn Close and musician Ozzy Osbourne reportedly did it to better understand mental illness in their families.
With genome sequencing costs plummeting from an initial $22 million to hovering around $1,000, it’s no longer restricted to the ultra wealthy, and in fact possibly five million human genomes will be sequenced by 2020. The Precision Medicine Initiative Cohort Program, which I blogged about two weeks ago, will contribute one million of them.
Meanwhile, genome (and exome, the protein-encoding part) sequencing is being used increasingly to end multi-year diagnostic odysseys, as I wrote about here at DNA Science, for JAMA, and for Medscape, which link to the centers doing the work. Knowing a sick child’s genome sequence can explain past symptoms while guiding future care and treatment choices. And it isn’t just sick kids. Human genomes are being sequenced from gametes to embryos to centenarians, the oldest old.
Of course genome sequencing extending back in time isn’t new, just on a different scale from The Genomic Postmortem Research Project’s 300 samples. The publication list of Svante Pääbo, Director of the Max Planck Institute for Evolutionary Anthropology, reveals these efforts. Since 1980, he’s led projects to probe the bones, hair, spit, and excrement of many species, analyzing the genomes of Denisovans, Neanderthals, a single individual who lived in what is now Romania some 40,000 years ago who had very recent Neanderthal forebears, Egyptian mummies, and Ötzi the ice man. The first ancient African genome – from circa 4,500 years ago in Ethiopia – was reported just a few weeks ago, from Ron Pinhas at University College Dublin and Marcos Gallego Llorente from the University of Cambridge. Analysis of ancient DNA has revealed the ills of past peoples, from infections and injuries to inherited diseases.
DNA from the more recent past has been scrutinized too, including posthumous diagnosis of Rett syndrome from a baby tooth and “rescue karyotyping” to reconstruct genetic clues to past pregnancy losses.
The Genomic Postmortem Research Project weds the looks backward using preserved body parts with the more systematic protocol of prospective genome investigations. Project leader is Murray Brilliant, Ph.D., and collaborators Simon Lin, M.D. and Min He, Ph.D., with funding from Complete Genomics. The goal is “to sequence the entire genome in 300 patients with long-term electronic medical records at Marshfield Clinic to determine (after the fact) if genomic knowledge could have positively influenced their medical care.”
It’s hard to know what, exactly, to look for among all those genomes, especially because we don’t know what all the 20,000 or so genes and their many variants do. But the Marshfield Clinic’s website and newsletters offer a compelling list to start:
- Chronic diseases (age-related macular degeneration, heart attack, type 2 diabetes, cataracts, and glaucoma)
- Cancers of the prostate, breast, and lung
- Susceptibility to bacterial infections and flu
- Fragile X syndrome premutation prevalence
- Iron metabolism
- Pharmacogenetics of metformin, statins, proton pump inhibitors
The researchers are sorting through 27+ million gene variants from the corpses, seeking those that could be used for disease prevention, detection, and personalized treatment and monitoring. Dr. He listed resources:
- SeqHBase, a “big data toolset” distinguishes de novo (new) mutations from inherited ones, and whether the variants are present in one copy (a heterozygote) or two copies (a homozygote or a compound heterozygote).
- ClinVar looks for disease or drug-associated gene variants.
- The ACMG (American College of American Genetics and Genomics) lists 56 mutations associated with 24 conditions, known simply as “the 56,” that it recommends be reported to patients if they show up as secondary findings, because they are “actionable.”
- NHGRI lists 112 actionable secondary findings from its exome sequencing project, including 52 of the ACMG 56.
The most daunting challenge right now is to classify gene variants – we can’t call them all mutations anymore due to negative connotations – as to how harmful evidence indicates they are. To make a very long story short, variants are classified as pathogenic, likely pathogenic, the dreaded “variant of uncertain significance,” likely benign, and benign. As the number of sequenced genomes climbs, the proportions will shift, with the numbers of VUS and the use of the word “likely” declining as pathogenic and benign designations rise. Uncertainty will certainly fade. Dr. He raced through some preliminary data that indicated a disturbing number of “variants of uncertain significance.” Then he described two interesting cases.
#1 A woman died from breast and ovarian cancer at age 59. She had no family history of cancer, and had never had a mammogram. Yet her genome sequencing revealed a rare pathogenic variant of BRCA1. “If the gene variant had been detected prior to her diagnoses, at an early age, it is possible that the cancers could have been predicted and/or prevented,” Dr. He said.
#2 A patient with deep vein thrombosis, a vascularized kidney cancer that had spread to the brain and lungs, and arteriovenous malformations (tangles of abnormally connected blood vessels), had hereditary hemorrhagic telangiectasia type 1 – a disorder of the vasculature. The diagnosis explained the symptoms, and would have suggested drug treatments old (such as thalidomide) and new (such as the anti-VEGF drug bevacizumab (Avastin).
Pharmacogenetic findings were intriguing – 281 of the 300 deceased individuals had meaningful variants among the 30 genes that the FDA recognizes as influencing the metabolism of more than 100 drugs. Examples include variants of CYP2C9 and VKORC1 and warfarin dose; DPYD and TPMT and the cancer drugs 6-mercaptopurine and irinotecan; CYP2D6 and tamoxifen and codeine; and a frighteningly long list of anti-depressants and anti-psychotics that are under or over active depending upon genotype. Several individuals had variants of SLC01B1 that increase risk of developing myopathy with use of the cholesterol-lowering drug simvastatin(Zocor).
If my grandpa would have had crippling muscle pain from Zocor and my doc was now trying to foist it on me, I’d want to know. But because the Genomic Postmortem Research Project is what it says – research – a protocol for contacting descendants who might want to know relevant genetic risk information about drugs and disease susceptibilities isn’t in place. The bioethicists will have to figure out how to coax informed consent from corpses.
Meanwhile, it seems pretty clear that when newborn genome sequencing becomes routine, DNA information may be the greatest gift from generations past.
Next Thursday I will be moderating a session on the Precision Medicine Initiative at the National Organization for Rare Disorders Summit in Washington, DC. I’ll be sure to eavesdrop again!
The NSA sure breaks a lot of "unbreakable" crypto. This is probably how they do it.
Sjon shared this story from Boing Boing. |
There have long been rumors, leaks, and statements about the NSA "breaking" crypto that is widely believed to be unbreakable, and over the years, there's been mounting evidence that in many cases, they can do just that. Now, Alex Halderman and Nadia Heninger, along with a dozen eminent cryptographers have presented a paper at the ACM Conference on Computer and Communications Security (a paper that won the ACM's prize for best paper at the conference) that advances a plausible theory as to what's going on. In some ways, it's very simple -- but it's also very, very dangerous, for all of us. (more…)
DNA.land: Enabling Crowd Sourced Genomic Research
DNA.land, a new webservice enabling crowdsourced genomic research, was launched by a team led by Yaniv Erlich and Joe Pickrell from NYGC at this year’s ASHG. The new site enables anyone to share their genetic data with an “easy to understand consent form”, learn more about their genome and help scientists use the genetic data for genomic research.
Thanks to low-cost microarray technologies, the personal genetic testing services like 23andMe and ancestry.com are helping thousands of individuals to get their genetic data and learn more about personal genetics. However, the genetic data are typically confined within such companies.
More recently, noble efforts like openSNP.org and Personal Genome Project, enable individuals to share their genetic data openly and freely to everyone. openSNP.org, in four years since opening, has over 4,000 individuals sharing genetic and phenotypic data openly. Similarly, Personal Genome Project has been making big strides in expanding to sequencing more individuals’ whole genome and making it and phenotype data available freely. In addition to sequence data, PGP is also sharing some data from services like 23andMe.
The number of people who have shared their genetic data openly to a project like openSNP is still small compared tot the number of people who have genetic data and also it is small for making new genetic discoveries. (Just earlier this summer 23andMe had its one millionth customer and the number of genotyped individuals will only increase in the coming years. ) One of the bottlenecks for seeing huge growth in adapting crowdsourced genomics is that, not all are comfortable with sharing the genetic data openly.
The new project, DNA.land, by academic scientists with clear consent and respect for privacy, is poised to make a difference in enabling crowdsourced genomic research. Within a week of DNA.land’s launch, over 5,600 individuals have shared their genetic data.
For now, DNA.land lets users upload the DNA data from services like 23andMe and infers ancestry, relatives based on the shared genomic segments. Within a few hours of data upload, DNA.land also offers imputed genome, imputed by using the reference population data. Pretty soon, DNA.land will ask the users to fill out surveys related to ancestry and health and use it to make discoveries.
If you have genotyped yourself, go ahead and sign up with DNA.land, learn more about your genome and enable science.
Also check out the nice piece on DNA.land at Nature Scientists hope to attract millions to ‘DNA.LAND’ (would have loved to see it mention openSNP, one of the early attempts at sharing genetic data.)
Video Tip of the Week: 100,000 Genomes Project
Siva, N. (2015) UK gears up to decode 100 000 genomes from NHS patients. The Lancet, 385(9963), 103-104. DOI: 10.1016/S0140-6736(14)62453-3
UK gears up to decode 100 000 genomes from NHS patients23andMe raises $115M to grow its DNA health test technology globally
23andMe has raised a substantial $115 million in a series E round led by Fidelity Management and Research Company, with participation from new investors Casdin Capital, WuXi Healthcare Ventures, and Xfund, as well as existing investors such as Google Ventures.
The personal genomics and biotechnology company was founded in 2006 by Linda Avey and Anne Wojcicki, wife of Google cofounder Sergey Brin. The Mountain View, Calif.-based company had already raised $190 million prior to today’s announcement, and boasts a slew of big-name investors, including Google, Google Ventures, Sergey Brin, and Yuri Milner.
23andMe, which is named for the 23 pairs of chromosomes found in a human cell, offers health tests directly to consumers. For $99 and a sample of DNA, the company will send you a report with information about your health and ancestry, covering 254 diseases and conditions.
In 2013, 23andMe received a public warning letter from the Food and Drug Administration (FDA) for marketing its genetic testing kit without proper clearance, though approval was formally granted earlier this year making it one of the first such companies to receive such approval. Today, it claims one million customers around the world.
The new funds will be used to “expand its operations and maximize the potential of its direct-to-consumer product in the U.S. and abroad,” 23andMe said in a press release. The company added that it will introduce a new user experience that includes carrier status reports by the end of 2015, alongside other new tools. The cash influx will also help it build new laboratory space for research.
Sweden is on track to becoming the first cashless nation
Improving YouTube video thumbnails with deep neural nets
Video thumbnails are often the first things viewers see when they look for something interesting to watch. A strong, vibrant, and relevant thumbnail draws attention, giving viewers a quick preview of the content of the video, and helps them to find content more easily. Better thumbnails lead to more clicks and views for video creators.
Inspired by the recent remarkable advances of deep neural networks (DNNs) in computer vision, such as image and video classification, our team has recently launched an improved automatic YouTube "thumbnailer" in order to help creators showcase their video content. Here is how it works.
The Thumbnailer Pipeline
While a video is being uploaded to YouTube, we first sample frames from the video at one frame per second. Each sampled frame is evaluated by a quality model and assigned a single quality score. The frames with the highest scores are selected, enhanced and rendered as thumbnails with different sizes and aspect ratios. Among all the components, the quality model is the most critical and turned out to be the most challenging to develop. In the latest version of the thumbnailer algorithm, we used a DNN for the quality model. So, what is the quality model measuring, and how is the score calculated?
The main processing pipeline of the thumbnailer. |
Unlike the task of identifying if a video contains your favorite animal, judging the visual quality of a video frame can be very subjective - people often have very different opinions and preferences when selecting frames as video thumbnails. One of the main challenges we faced was how to collect a large set of well-annotated training examples to feed into our neural network. Fortunately, on YouTube, in addition to having algorithmically generated thumbnails, many YouTube videos also come with carefully designed custom thumbnails uploaded by creators. Those thumbnails are typically well framed, in-focus, and center on a specific subject (e.g. the main character in the video). We consider these custom thumbnails from popular videos as positive (high-quality) examples, and randomly selected video frames as negative (low-quality) examples. Some examples of the training images are shown below.
Example training images. |
Results
Compared to the previous automatically generated thumbnails, the DNN-powered model is able to select frames with much better quality. In a human evaluation, the thumbnails produced by our new models are preferred to those from the previous thumbnailer in more than 65% of side-by-side ratings. Here are some examples of how the new quality model performs on YouTube videos:
Example frames with low and high quality score from the DNN quality model, from video “Grand Canyon Rock Squirrel”. |
Thumbnails generated by old vs. new thumbnailer algorithm. |
Is Google a competitor or partner to Big Pharma? Roche's CEO doesn't know, yet
The advance of Google and other tech firms into life sciences has given drug developers a new source of partners with complementary skills but also raised an awkward question: Long term, are these companies allies or rivals? For Roche CEO Severin Schwan, the answer is still unclear.
PacBio Announces A New Sequencer: PacBio Sequel
In case you missed it, Last week, PacBio – the leader in long-read sequencing technology, announced that it is launching a new sequencer Sequel. Announcing the Sequel, just a few days before the start of this year’s ASHG conference at Baltimore, PacBio said it will showcase Sequel at the 2015 ASHG meeting. The Sequel system is expected to be shipped to limited customers, including Roche for its assay development program, during the 4th quarter of 2015. Shipments outside the U.S. are expected to commence thereafter.
In the press release announcing the Sequel, PacBio said
The Sequel System has been developed as part of the company’s collaboration with F. Hoffman-La Roche Ltd(Roche) to ultimately provide a nucleic acid sequencing system for use in human in vitro diagnostics.
Under the Roche-PacBio agreement, Roche will be paying $40M for the development of Sequel. Roche has already paid $20M and the remaining 20M is expected by the fourth quarter of 2015.
The Sequel vs PacBio RS II
The new sequencer Sequel is powered by PacBio’s redesigned SMRT cells, which contains one million zero-mode waveguides during the launch. In comparison to PacBio RS II system, The Sequel is a smaller machine with higher throughput. The new sequencer, Sequel is less than one-third the size and weight, when compared with PacBio RS II. And the Sequel has 1M ZMW, in comparison 150,000 ZMWs in the PacBio RS II system. Therefore, the through-put of the Sequel is 7 times more reads than PacBio RS II. The list price for the PacBio’s Sequel System in US is $350,000, in comparison to ~$750,000 for PacBio RS II.
With Sequel, one can get 10X coverage human data good enough to look at structural variants approximately about $3,000. For sequencing a human genome at 30X coverage, approximate cost is expected to be just around $10,000, in comparison to 20-25k using PacBio RS II.
The announcement of Sequel has been received very positively and has put PacBio in a strong position. The street reacted sharply with a spike in PacBio’s stocks and a drop in Illumina’s stock.
Here is the link-round up of blogs and press covering the launch of Sequel.
- PacBio Press Release: Pacific Biosciences Launches New Sequencing Platform Based on Its SMRT Technology
- PacBio Blog: Introducing the Sequel System: The Scalable Platform for SMRT Sequencing
- GenomeWeb: PacBio Launches Higher-Throughput, Lower-Cost Single-Molecule Sequencing System
- Bio IT: PacBio Announces Sequel Sequencing System
- Bio IT: A Worthy Sequel: PacBio’s New Sequencing System
- Bio Mick Watson: What does the PacBio Sequel mean for the future of sequencing?
- Core genomics: The new Pacific Biosciences sequencer
- Keith Robinson OmicsOmics Blog: PacBio Sequel: Smaller Box, Bigger Bang
- Mark Chaisson: What does SEQUEL mean for human genetics?
PacBio Sequel Introduction Video
PacBio Sequel on NBC TV
Twin Sons Of Different Mothers…… Or Fathers
Pollux, B., Meredith, R., Springer, M., Garland, T., & Reznick, D. (2014) The evolution of the placenta drives a shift in sexual selection in livebearing fish. Nature, 513(7517), 233-236. DOI: 10.1038/nature13451
The evolution of the placenta drives a shift in sexual selection in livebearing fishPeigné, M., Andrieux, J., Deruelle, P., Vuillaume, I., & Leroy, M. (2011) Quintuplets after a transfer of two embryos following in vitro fertilization: a proved superfecundation. Fertility and Sterility, 95(6), 2147483647-2147483647. DOI: 10.1016/j.fertnstert.2011.01.029
Quintuplets after a transfer of two embryos following in vitro fertilization: a proved superfecundationJames WH. (1993) The incidence of superfecundation and of double paternity in the general population. Acta geneticae medicae et gemellologiae, 42(3-4), 257-62. PMID: 7871943
The incidence of superfecundation and of double paternity in the general population.Baijal N, Sahni M, Verma N, Kumar A, Parkhe N, & Puliyel JM. (2007) Discordant twins with the smaller baby appropriate for gestational age--unusual manifestation of superfoetation: a case report. BMC pediatrics, 2. PMID: 17239246
Discordant twins with the smaller baby appropriate for gestational age--unusual manifestation of superfoetation: a case report.Claas, M., Timmermans, A., & Bruinse, H. (2010) Case report: a black and white twin. Journal of Perinatology, 30(6), 434-436. DOI: 10.1038/jp.2009.156
Case report: a black and white twinPacBio soars on news of cheaper, smaller, higher-throughput sequencer
Pacific Biosciences has stepped up the fight for the sequencing market with the introduction of an instrument that is smaller, cheaper and higher throughput than the model it replaces.