DOE Joint Genome Institute

  • COVID-19
  • About
  • Phones
  • Contacts
  • Our Science
    • DOE Mission Areas
    • Bioenergy Research Centers
    • Science Programs
    • Products
    • Science Highlights
    • Scientists
    Maize can produce a cocktail of antibiotics with a handful of enzymes. (Sam Fentress, CC BY-SA 2.0)
    How Maize Makes An Antibiotic Cocktail
    Zealexins are produced in every corn variety and protect maize by fending off fungal and microbial infections using surprisingly few enzymes.

    More

    The genome of the common fiber vase or Thelephora terrestris was among those used in the study. (Francis Martin)
    From Competition to Cooperation
    By comparing 135 fungal sequenced genomes, researchers were able to carry out a broader analysis than had ever been done before to look at how saprotrophs have transitioned to the symbiotic lifestyle.

    More

    Miscanthus grasses. (Roy Kaltschmidt/Berkeley Lab)
    A Grass Model to Help Improve Giant Miscanthus
    The reference genome for M. sinensis, and the associated genomic tools, allows Miscanthus to both inform and benefit from breeding programs of related candidate bioenergy feedstock crops such as sugarcane and sorghum.

    More

  • Our Projects
    • Search JGI Projects
    • DOE Metrics/Statistics
    • Approved User Proposals
    • Legacy Projects
    Poplar (Populus trichocarpa and P. deltoides) grow in the Advanced Plant Phenotyping Laboratory (APPL) at Oak Ridge National Laboratory in Tennessee. Poplar is an important biofuel feedstock, and Populus trichocarpa is the first tree species to have its genome sequenced — a feat accomplished by JGI. (Image courtesy of Oak Ridge National Laboratory, U.S. Dept. of Energy)
    Podcast: Xiaohan Yang on A Plantiful Future
    Building off plant genomics collaborations between the JGI and Oak Ridge National Laboratory, Xiaohan Yang envisions customizing plants for the benefit of human society.

    More:

    Expansin complex with cell wall in background. (Courtesy of Daniel Cosgrove)
    Synthesizing Microbial Expansins with Unusual Activities
    Expansin proteins from diverse microbes have potential uses in deconstructing lignocellulosic biomass for conversion to renewable biofuels, nanocellulosic fibers, and commodity biochemicals.

    Read more

    High oleic pennycress. (Courtesy of Ratan Chopra)
    Pennycress – A Solution for Global Food Security, Renewable Energy and Ecosystem Benefits
    Pennycress (Thlaspi arvense) is under development as a winter annual oilseed bioenergy crop. It could produce up to 3 billion gallons of seed oil annually while reducing soil erosion and fertilizer runoff.

    Read more

  • Data & Tools
    • IMG
    • Genome Portal
    • MycoCosm
    • PhycoCosm
    • Phytozome
    • GOLD
    Artistic interpretation of CheckV assessing virus genome sequences from environmental samples. (Rendered by Zosia Rostomian​, Berkeley Lab)
    An Automated Tool for Assessing Virus Data Quality
    CheckV can be broadly utilized by the research community to gauge virus data quality and will help researchers to follow best practices and guidelines for providing the minimum amount of information for an uncultivated virus genome.

    More

    Unicellular algae in the Chlorella genus, magnified 1300x. (Andrei Savitsky)
    A One-Stop Shop for Analyzing Algal Genomes
    The PhycoCosm data portal is an interactive browser that allows algal scientists and enthusiasts to look deep into more than 100 algal genomes, compare them, and visualize supporting experimental data.

    More

    Artistic interpretation of how microbial genome sequences from the GEM catalog can help fill in gaps of knowledge about the microbes that play key roles in the Earth's microbiomes. (Rendered by Zosia Rostomian​, Berkeley Lab)
    Podcast: A Primer on Genome Mining
    In Natural Prodcast: the basics of genome mining, and how JGI researchers conducted it in IMG/ABC on thousands of metagenome-derived genomes for a Nature Biotechnology paper.

    Read more

  • User Programs
    • Calls for User Proposals
    • Special Initiatives & Programs
    • User Support
    • Submit a Proposal
    Scanning electron micrographs of diverse diatoms. (Credits: Diana Sarno, Marina Montresor, Nicole Poulsen, Gerhard Dieckmann)
    Learn About the Approved 2021 Large-Scale CSP Proposals
    A total of 27 proposals have been approved through JGI's annual Community Science Program (CSP) call. For the first time, 63 percent of the accepted proposals come from researchers who have not previously been a principal investigator on an approved JGI proposal.

    Read more

    MiddleGaylor Michael Beman UC Merced
    How to Successfully Apply for a CSP Proposal
    Reach out to JGI staff for feedback before submitting a proposal. Be sure to describe in detail what you will do with the data.

    Read more

    Click on the image or go here to watch the video "Enriching target populations for genomic analyses using HCR-FISH" from the journal Microbiome describing the research.
    How to Target a Microbial Needle within a Community Haystack
    Enabled by the JGI’s Emerging Technologies Opportunity Program, researchers have developed, tested and deployed a pipeline to first target cells from communities of uncultivated microbes, and then efficiently retrieve and characterize their genomes.

    Read more

  • News & Publications
    • News
    • Blog
    • Podcasts
    • Publications
    • Scientific Posters
    • Newsletter
    • Logos and Templates
    • Photos
    Artistic interpretation of how microbial genome sequences from the GEM catalog can help fill in gaps of knowledge about the microbes that play key roles in the Earth's microbiomes. (Rendered by Zosia Rostomian​, Berkeley Lab)
    Uncovering Novel Genomes from Earth’s Microbiomes
    A public repository of 52,515 microbial draft genomes generated from environmental samples around the world, expanding the known diversity of bacteria and archaea by 44%, is now available .

    More

    Green millet (Setaria viridis) plant collected in the wild. (Courtesy of the Kellogg lab)
    Shattering Expectations: Novel Seed Dispersal Gene Found in Green Millet
    In Nature Biotechnology, a very high quality reference Setaria viridis genome was sequenced, and for the first time in wild populations, a gene related to seed dispersal was identified.

    More

    The Brachypodium distachyon-B. stacei-B. hybridum polyploid model complex. (Illustrations credits: Juan Luis Castillo)
    The More the Merrier: Making the Case for Plant Pan-genomes
    Crop breeders have harnessed polyploidy to increase fruit and flower size, and confer stress tolerance traits. Using a Brachypodium model system, researchers have sought to learn the origins, evolution and development of plant polyploids. The work recently appeared in Nature Communications.

    Read more

News & Publications
Home › Blog › Accelerating Metagenomics Discoveries with Deep Learning

July 28, 2020

Accelerating Metagenomics Discoveries with Deep Learning

An example protein structure: the orange carotenoid protein. Proteins would be easier to align if they could be vectorized, acccording to Sevim. (Kerfeld lab group)
An example protein structure: the orange carotenoid protein. Proteins would be easier to align if they could be vectorized, according to JGI scientist Volkan Sevim. (Kerfeld lab group)

On July 6, scientists from Lawrence Berkeley National Laboratory (Berkeley Lab) and other institutions shared their works-in-progress with 86 participants in an online workshop to harness deep learning for metagenomic data. 

Metagenomic data, or DNA sequencing data gathered from whole samples rather than single cells or isolated strains, is easy to generate, but difficult to assemble and interpret. The workshop aimed to showcase how researchers are utilizing deep learning to automatically generate insights from metagenomic data.

In deep learning, a branch of machine learning, data is passed through algorithms organized into multiple interconnected layers. The architecture of these algorithms, inspired by communications networks found in nature, allows machines to “learn” features. State of the art algorithms can perform astonishing degrees of feature recognition: for example, recognizing your face and distinguishing it from your friend’s. Biologists aim to capitalize on these kinds of abilities by applying them to sequencing data.

Volkan Sevim, a computational biologist at the US Department of Energy Joint Genome Institute (JGI), presented his research to create a more efficient protein search tool using deep learning. (Photo courtesy of Volkan Sevim)
Volkan Sevim, a computational biologist at the US Department of Energy Joint Genome Institute (JGI), presented his research to create a more efficient protein search tool using deep learning. (Photo courtesy of Volkan Sevim)

At the workshop, Volkan Sevim, a computational biologist at the US Department of Energy (DOE) Joint Genome Institute (JGI), presented on generating vector representations of proteins in order to conduct more efficient protein sequence alignments. He also aims to detect structural similarities in proteins that low sequence identity may belie. The technology he’s using is magical; transform any length of amino acid sequence into a smaller, constant-length vector using deep learning. Vector representations allow homology searches between proteins to become much simpler; instead of aligning sequences with popular, but computationally costly approaches such as BLAST and hidden Markov models (HMMs), “You just compute the distance between two vectors,” said Sevim. 

While there are alignment-free methods for nucleotide searches, such as Mash and BBsketch, whose lead developer is JGI software developer Brian Bushnell, no alignment-free tool exists for proteins. (And while Mash and BBsketch can be used on nucleotide sequences, they’re not as reliable on amino acid sequences.) Sevim is currently using HMM profiles of protein domains from the pfam database to test the efficacy of a particular kind of deep learning algorithm — a convolutional neural network, called ResNet 34  — in order to vectorize protein sequences.

Sevim is part of the JGI Genome R&D Group and is one of more than 100 scientists and systems engineers involved in computational work at the JGI. Sevim’s work is part of an ongoing effort to build greater collaboration between JGI and scientists in the Computational Resources Division (CRD), in order to create more useful tools for the bioinformatics community.

Other work presented at the workshop included how to use deep learning to predict protein function (Héctor García Martin of Berkeley Lab’s Biological Systems and Engineering), to assign microbial taxa (Andrew Tritt of CRD and, independently, Ying Zhang of the University of Rhode Island), and to bin metagenomic contigs (inferred chromosomal sequences in mixed-genome samples) with a new tool called MetaGNN (Prashant Pandey of CRD).

The workshop was organized by Kristofer Bouchard, neuro- and data scientist in Biological Systems and Engineering / Computing Resources Division (BSE/CRD).

Deep learning algorithms can vectorize faces...
Deep learning algorithms can vectorize faces…
... so, why not proteins? (Screenshots from Sevim's online presentation, Alison F. Takemura)
… so, why not proteins? (Screenshots from Sevim’s online presentation, Alison F. Takemura)

By Alison F. Takemura

Share this:

  • Click to share on Facebook (Opens in new window)
  • Click to share on LinkedIn (Opens in new window)
  • Click to share on Pinterest (Opens in new window)
  • Click to share on Twitter (Opens in new window)
  • Click to print (Opens in new window)

Filed Under: Blog

More topics:

  • COVID-19 Status
  • News
  • Science Highlights
  • Blog
  • Podcasts
  • CSP Plans
  • Featured Profiles

Related Content:

JGI on the 2020 Highly Cited Researchers List

2020 Highly Cited Researchers at the JGI

Tanja Woyke Awarded van Niel International Prize for Studies in Bacterial Systematics

Tanja Woyke JGI

Harnessing JGI’s Metabolomics Capabilities

JGI engagement webinar:harnessing metabolomics capabilities

UC Merced Interns Reflect on Their JGI Summer Projects

Screencap Axel Visel intro video

JGI Welcomes New UEC Members

(left to right) Kathleen Greenham of the University of Minnesota, Matthias Hess of the University of California, Davis and Kristen DeAngelis of University of Massachusetts-Amherst,

The JGI Data Portal: Improving User Experience

JGI Data Portal screencap
  • Careers
  • Contact Us
  • Events
  • User Meeting
  • MGM Workshops
  • Internal
  • Disclaimer
  • Credits
  • Emergency Info
  • Accessibility / Section 508 Statement
  • RSS feed
  • Flickr
  • LinkedIn
  • Twitter
  • YouTube
Lawrence Berkeley National Lab Biosciences Area
A project of the US Department of Energy, Office of Science

JGI is a DOE Office of Science User Facility managed by Lawrence Berkeley National Laboratory

© 1997-2021 The Regents of the University of California