<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Lifebit - Medium]]></title>
        <description><![CDATA[Building the world&#39;s first intelligent genomics platform that understands DNA data and generates meaningful insights like humans do. - Medium]]></description>
        <link>https://medium.com/lifebit-ai?source=rss----34b6fcaa2544---4</link>
        <image>
            <url>https://cdn-images-1.medium.com/proxy/1*TGH72Nnw24QL3iV9IOm4VA.png</url>
            <title>Lifebit - Medium</title>
            <link>https://medium.com/lifebit-ai?source=rss----34b6fcaa2544---4</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sun, 22 Nov 2026 04:15:24 GMT</lastBuildDate>
        <atom:link href="https://medium.com/feed/lifebit-ai" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Bringing together high-quality open-source community resources with Deploit]]></title>
            <link>https://medium.com/lifebit-ai/bringing-together-high-quality-open-source-community-resources-with-deploit-6a32d34bf1de?source=rss----34b6fcaa2544---4</link>
            <guid isPermaLink="false">https://medium.com/p/6a32d34bf1de</guid>
            <category><![CDATA[cloud-computing]]></category>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[genomics]]></category>
            <category><![CDATA[open-source]]></category>
            <category><![CDATA[bioinformatics]]></category>
            <dc:creator><![CDATA[Luisa Santus]]></dc:creator>
            <pubDate>Thu, 05 Jul 2018 16:25:50 GMT</pubDate>
            <atom:updated>2018-07-05T16:29:15.408Z</atom:updated>
            <content:encoded><![CDATA[<h4>A Variant Calling use-case.</h4><h3>Challenges in a Bioinformatics Open-Source Era</h3><p>Over the last few years open-source has been increasingly becoming the norm, even in Bioinformatics. The number of high-quality applications which are <a href="https://github.com/gatk-workflows/">freely available on GitHub</a> and other Git providers is increasing, such as the pipelines that the <a href="https://www.broadinstitute.org/">Broad institute</a> uses for production.</p><p>On the other hand, there has been considerable effort to collect publicly available datasets, across many different fields in biology/bioinformatics. The list is long but you can find one <a href="https://en.wikipedia.org/wiki/List_of_biological_databases">here</a> of some databases containing data related to bioinformatics. Finally, some cloud providers are also putting at disposal biological datasets of interest, like AWS does with the <a href="https://aws.amazon.com/it/opendata/">Open Data</a> project.</p><p><strong>All of this is thrilling</strong> — such great resources becoming freely available gives room to a lot of new research and applications!</p><p><strong>Although the potential is there, the real revolution will come when all of this can be brought together</strong> and produce insights that drive us forward.</p><blockquote><strong>Although the potential is there, the real revolution will come when all of this can be brought together</strong> and produce insights that drive us forward.</blockquote><p>The challenge is that bringing it all together, most of the time, is not easy. Data lives on different cloud storage providers (AWS S3 storage, Azure blob storage etc.), public available databases or on-premise storage systems. Pipelines can be found on different code versioning platforms (GitHub, Bitbucket etc.) or different container hubs (<a href="https://hub.docker.com/">DockerHub</a>, <a href="https://quay.io/search">Quay</a> etc.). Lastly, various on-premise high performance systems or cloud providers are used for computational power.</p><p>So when we actually run analyses, we first have to collect everything we need and handle compatibility, accessibility and quality assurance of the collected resources before we can begin the analyses — which is the core part of our work.</p><h3>Deploit as the Connection Centre</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*R5xZ65aVM16jWrzylZqgrg.png" /></figure><p>In the currently growing bioinformatics open-source era, Deploit brings together data, applications and resources, which would otherwise be dispersed. Deploit synchronizes with all the major software development tools and cloud providers. This enables easy access and running of open-source pipelines on publicly available data on the cloud, no matter where the pipeline or data are stored or which cloud provider you use for computation. All of it has been packaged in a simple and frictionless user experience and interface that also enables people with minimal to zero computational training to run and scale analyses over cloud.</p><blockquote>Deploit brings together data, applications and resources</blockquote><p>At the moment, we are testing and making available a collection of high-quality pipelines that are popular among the community, off-the-shelf on the platform. We have already added a large list of popular <a href="https://github.com/nextflow-io/awesome-nextflow">Nextflow</a> and Docker pipelines, and we will keep expanding it with more pipelines (Snakemake and CWL pipelines coming soon).</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/600/0*MQvBTSpSz8wgNz2Q.gif" /></figure><p>Moreover, we support import and visualization of data from major cloud providers. Computation can be triggered through Deploit on different cloud providers as well.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*5kxO2F2Kf26lXbWS.png" /></figure><h3>A Variant Calling Use-case</h3><p><a href="https://www.ebi.ac.uk/training/online/course/human-genetic-variation-i-introduction/variant-identification-and-analysis/what-variant">Variant Calling</a> is a widely used tool, which identifies variants from sequence data. In this use-case it will be shown how to run <a href="https://github.com/lifebit-ai/freebayes">FreeBayes as a Nextflow pipeline</a> ( an open-source variant caller) on publicly available data over cloud with Deploit.</p><h3>Open-Source Pipeline: FreeBayes</h3><p>For this example, we are going to use FreeBayes as a pipeline. In Deploit, we also provide many other variant callers like <a href="https://github.com/lifebit-ai/DeepVariant">DeepVariant</a> and <a href="https://github.com/lifebit-ai/strelka2">Strelka2</a> as Nextflow pipelines, as well as easy access to many other open-source pipelines.</p><h3>Public Dataset: Genome in a Bottle!</h3><p>Although the quality of the developed methods is advancing, different variant callers use different algorithms so may still disagree on their predictions. This is when benchmarking becomes essential to assess the quality of the predictions produced in order to identify those that are meaningful for clinical practice. The <a href="http://jimb.stanford.edu/giab/">Genome in a Bottle Consortium</a> has <a href="https://www.nature.com/articles/nbt.2835">released</a> a set of high-confidence, experimentally validated set of variants publicly available in a <a href="https://s3.console.aws.amazon.com/s3/buckets/giab">S3 bucket</a>. This public reference material is extremely precious and various studies and benchmarks have already been conducted based on these datasets (e.g. <a href="https://www.ncbi.nlm.nih.gov/pubmed/25711446?dopt=Abstract&amp;holding=npg">1</a>,<a href="https://www.hindawi.com/journals/bmri/2015/456479/">2</a>,<a href="https://www.nature.com/articles/srep17875#ref19">3</a>).</p><h3>Bringing it all together with Deploit</h3><p>One of Deploit’s strengths is the <strong>synchronization and integration capability</strong>. In fact, it synchronizes with the main cloud providers as well as software development tools.</p><p>In practical terms, this translates into a seamless and effortless user experience, enabling the user to create as many collections of datasets and pipelines as they wish in virtually zero time and cost.</p><p><strong>Porting a public or private pipeline</strong> in Deploit becomes as easy as providing and URL (copy-pasting) from GitHub, BitBucket or DockerHub.</p><p><strong>Accessing desired data</strong> can be achieved by either a simple upload of the data or a synchronization with private or public cloud storage. Deploit employs an intelligent virtual filesystem that allows the data to be linked but never moved or copied to the user’s cloud storage. It also prevents duplications of the data that already live within the user’s cloud storage. This minimises cost and data accessibility and management time.</p><p>Through a synchronization to your cloud account, Deploit manages the computation and delivers to the end user the final results.</p><p><strong>Let’s have a concrete look on how it works</strong> through an example: running an open-source variant caller on an open source dataset over cloud in a scalable and reproducible manner.</p><p>In order to do so with Deploit, one of the featured variant callers can be selected: e.g FreeBayes. Afterwards, the content of the GIAB S3 bucket , as well an any other public S3 bucket, can be browsed, and bam files, the input required by FreeBayes, can be selected. Eventually, the computation is triggered by selecting the resource configuration. Done!</p><p>By the way: we are now on open-beta → <a href="http://deploit.lifebit.ai/register/">sign up</a> and give it a try!</p><p>In the following gif you can observe how the whole journey looks like in Deploit.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/600/0*LaMlC199FKfT0BeY.gif" /></figure><p><strong>We are still at the beginning and we have big plans for the future! Don’t miss a thing and let us keep you updated by following us on </strong><a href="https://twitter.com/Lifebit_Biotech"><strong>Twitter</strong></a><strong>.</strong></p><p><em>We’re also actively looking for great engineers and bioinformaticians who want to help us shape Lifebit — if this is something you’d be interested in, we’d love to hear from you: hello@lifebit.ai.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=6a32d34bf1de" width="1" height="1" alt=""><hr><p><a href="https://medium.com/lifebit-ai/bringing-together-high-quality-open-source-community-resources-with-deploit-6a32d34bf1de">Bringing together high-quality open-source community resources with Deploit</a> was originally published in <a href="https://medium.com/lifebit-ai">Lifebit</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[DeepVariant as a Nextflow Pipeline on Deploit]]></title>
            <link>https://medium.com/lifebit-ai/deepvariant-as-a-nextflow-pipeline-on-deploit-4066e572eab8?source=rss----34b6fcaa2544---4</link>
            <guid isPermaLink="false">https://medium.com/p/4066e572eab8</guid>
            <category><![CDATA[deep-learning]]></category>
            <category><![CDATA[google]]></category>
            <category><![CDATA[genomics]]></category>
            <category><![CDATA[genetics]]></category>
            <category><![CDATA[cloud-computing]]></category>
            <dc:creator><![CDATA[Luisa Santus]]></dc:creator>
            <pubDate>Tue, 19 Jun 2018 15:41:51 GMT</pubDate>
            <atom:updated>2018-07-05T16:29:51.757Z</atom:updated>
            <content:encoded><![CDATA[<blockquote>How to run DeepVariant on Deploit in a reproducible, scalable and parallelised manner over cloud.</blockquote><p>The beginning of the 21st century has been shaped by the fast growing amount of biological data which is produced by sequencing technologies and, in particular, by the high throughput next generation sequencing<strong> </strong>(NGS). Genetic variations in human genomes, especially the very common single nucleotide polymorphisms (<strong>SNPs</strong>), are correlated with many diseases [1], associated with individuality [2] and relevant to many other fields, such as nutrition [3]. In the last few years, interest in genomic variations and their effects on phenotype and impact on diseases has been growing. The first step to harness the genotype-phenotype connection is to be able, given an individual, to identify the variations in his/her genome.</p><p><a href="https://www.ebi.ac.uk/training/online/course/human-genetic-variation-i-introduction/variant-identification-and-analysis/what-variant"><strong>Variant Calling</strong></a> tools identify genetic variations of newly sequenced genomes by comparing to the corresponding reference genome. In practice, the output of NGS technologies (reads) are mapped to the reference genome and variations are identified. In the last years, many variant callers have been released in order to obtain accurate and fast results. In December 2017, the team at Google Brain joined the effort by releasing an open source, <strong>deep learning based variant caller</strong>: <a href="https://github.com/google/deepvariant">DeepVariant</a>. DeepVariant outperforms its competitors by accuracy — it even won the accuracy award at the <a href="https://precision.fda.gov/challenges/truth/results">precisionFDA Truth Challenge</a>.</p><p>As we started using DeepVariant, we realized that running the binaries provided by Google over our datasets was not efficient enough. Parallelization over multiple files is not handled by default, input files need some preprocessing since not all of them are compatible with DeepVariant, and manually running the binaries or, writing some temporary script on top of them, was not sustainable. Moreover, we needed a way to seamlessly run DeepVariant over cloud in the easiest way possible. That’s when we decided to write a <strong>Nextflow pipeline and use it on Deploit </strong>to solve our problems.</p><p><a href="https://github.com/lifebit-ai/DeepVariant"><strong>DeepVariant as a Nextflow pipeline</strong></a><strong> </strong>enables users to run DeepVariant in an easy, fast and reproducible manner that ensures <strong>full control</strong> over configurations. It supports the user to start running the pipeline correctly and it allows running of multiple variant calling processes in parallel to maximize efficiency. Every step is bundled and computed in a Docker container. Full support to easily run and scale analyses with DeepVariant over AWS and Azure cloud is provided by Deploit.</p><h3><strong>DeepVariant: an overview</strong></h3><p>DeepVariant performs variant calling through <strong>image recognition</strong>. Images representing the mapping of the reads to the reference genome are produced using BAM files. The picture below, taken from the official <a href="https://research.googleblog.com/2017/12/deepvariant-highly-accurate-genomes.html">deep variant google blog post</a>, shows examples of such images. These images are then used by the trained machine learning model to identify the variants.</p><p>DeepVariant’s <a href="https://github.com/google/deepvariant">release</a> workflow is divided in three different binaries.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/742/0*FESnCm7mq6I3SPg-" /></figure><p><strong><em>Make_examples: </em></strong>The make_example step uses the reference genome and the <a href="https://genome.sph.umich.edu/wiki/BAM">BAM file</a> in order to <strong>produce the piled-up images</strong> needed for the prediction. To speed things up, the make_example step can be parallelized and take advantage of multiple machine’s cores.</p><p><strong><em>Call_variants: </em></strong>The call_variants step performs the real variant calling: it uses the already trained prediction model to identify genomic variants.</p><p><strong><em>Postprocess_variants: </em></strong>This step only converts the output of call variants in the standard and well-known <a href="http://www.internationalgenome.org/wiki/Analysis/vcf4.0/">VCF </a>format.</p><p>A more detailed description of the steps can be found <a href="https://github.com/google/deepvariant">here</a>.</p><h3><strong>DeepVariant in Nextflow</strong></h3><p>As shown in the following picture, the Nextflow version of the DeepVariant workflow contains not only the<strong> variant calling steps </strong>described above (darker blue ones) but also some <strong>preprocessing steps</strong> (light blue ones) as well.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/501/0*t2iuSQCDBsQvPz18" /></figure><p>This allows for the whole analysis to run seamlessly and reproducibly instead of running manually 3 binaries with all the irreproducible mess, ineffectiveness and frustration that this may cause.</p><p>The workflow handles one reference genome and <strong>multiple BAM files</strong> as input. The variant calling for the several input BAM files will be processed completely independently and will produce independent VCF result files. The minimum set of <strong>input parameters</strong> is only composed by the version of the reference genome (available on Lifebit) and a folder where your bam files are stored. When running the pipeline in this way, the other needed files will be automatically created and the last release of google’s trained model for whole genome variant calling will be used.</p><p>It looks as simple as this:</p><pre>nextflow run main.nf --hg19 --bam_folder “path/to/folder/”</pre><p>Despite the possibility of <strong>keeping things this simple</strong>, complete <strong>control</strong> over all the configurations is ensured. Calling the pipeline can be as flexible as this:</p><pre>nextflow run main.nf — fasta “path/to/my/genome.fasta” <br> — fai “path/to/my/genome.fasta.fai”<br> — fastagz “path/to/my/genome.fasta.gz”<br> — gzfai “path/to/my/genome.fasta.gz.fai”<br> — gzi “path/to/my/genome.fasta.gz.gzi”<br> — bam_folder “path/to/folder/with/bam/files”<br> — getBai “true”<br> — j 64<br> — modelFolder “path/to/the/folder/with/my/model/files”<br> — modelName “model.ckpt”</pre><p>It allows users to pass all the input files (compressed and indexed), define how many cores should be used for the parallelization and even change the model which should be used. In this way, users can <strong>define their level of control</strong>.</p><p>Finally, the advantage of this approach is that the variant calling of the different BAM files can be parallelized internally by Nextflow and all cores of the machine are taken advantage of in order to get to the results faster.</p><h3><strong>DeepVariant at its best: Nextflow &amp; Lifebit Deploit</strong></h3><p><strong>Running pipelines over the cloud can be a struggle</strong> and make you lose precious time if you are not an experienced cloud user: it begins with knowing how to fire up an instance and pick a suitable one, then how to transfer your data, how to clone your repositories, install all the dependencies and finally remembering to terminate the machines when the job is done to stay on budget. Honestly, this is not fun!! Which made us wonder in despair: <strong>why can’t the cloud be a simple compute tool that anyone could use without being cloud experts?</strong> We should not be losing time figuring out how the cloud works but rather concentrating on the results we are trying to obtain.</p><blockquote><strong>Why can’t the cloud be a simple compute tool that anyone could use without being cloud experts?</strong></blockquote><p><strong>We built Deploit to help relieve this pain</strong> — it saves time and helps you pick the fastest and most cost-effective configuration to run and scale your analysis over cloud.</p><p>It allows you, in a matter of a couple of clicks, to synchronize your cloud account as well as your GitHub or Bitbucket account. You can then run <em>your</em> pipeline on <em>your</em> data in a scalable manner over the cloud and still get some sleep at night!</p><p>Deploit brings the<strong> easiness of running DeepVariant</strong> to a whole new level by providing a nice user interface which can guide the user through the journey of successfully and efficiently deploying and running variant calling analyses using DeepVariant over the cloud.</p><h3><strong>Epilogue on runtimes and costs</strong></h3><p>Using <strong>DeepVariant through Deploit</strong> not only makes life easier, it actually <strong>speeds up running times</strong> and automatically makes the most out of the given resources.</p><p>We <strong>benchmarked</strong> DeepVariant running times over 10 different BAM files from <a href="http://hgdownload.cse.ucsc.edu/goldenPath/hg19/encodeDCC/wgEncodeUwRepliSeq/">UCSC</a> on 3 different AWS machines with different resources (<em>m4.16xlarge, cr1.8xlarge, c3.4xlarge</em>).</p><p>We ran the binaries manually as described in the <a href="https://github.com/google/deepvariant/blob/r0.5/docs/deepvariant-docker.md">google DeepVariant documentation</a> taking care manually of preprocessing, transfering of data and cloud configurations (labelled “<em>Bash Script</em>” below) and DeepVariant as a Nextflow pipeline on Deploit (labelled “<em>Resource optimized with Deploit</em>” below).</p><p>Results show that <strong>DeepVariant in Nextflow on Deploit outperforms the manual (bash script) version</strong> of DeepVariant across both time and costs. While looking at this results one should also take into account that, as a user, running DeepVariant on Deploit took almost no time: you simply add input parameters and it’s done! On the other hand, <strong>running it manually needs time</strong> and effort, from generating all the needed input files (automatically handled in the Nextflow pipeline), manually moving data back and forth from Docker containers and having to trigger the computation multiple times in order to run DeepVariant on multiple BAM files. <em>(NOTE: the time needed by these steps is NOT included in the benchmark, where only runtimes are shown)</em>. Not only are DeepVariant runtimes much lower by using Deploit, but also the time needed for setup is eliminated.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*xPTqJmQjFciw4TvR" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*mcZr5FoSMNZ0ATQ_" /></figure><p>As it can be observed in the benchmarking graphs, Deploit can noticeably reduce time and lower costs. Time decrease is achieved by carefully using auto scaling and parallelization while low costs are ensured by Deploit’ resource optimization based on spot instances.</p><p>All data and material used for this benchmark can be found in this publicly available <a href="https://s3.console.aws.amazon.com/s3/#">s3 bucket</a>. <em>(The prices used for the benchmarking are based on the 16th June 2018 instances prices).</em></p><p><em>If you are interested in what we are doing and want to have early access to the platform please feel free to contact us at hello@lifebit.ai.</em></p><p><em>We are still at the beginning and we have big plans for the future! Don’t miss a thing and let us keep you updated on our following posts by signing up to our Newsletter or visit us on Twitter</em></p><p><em>We’re also actively looking for great engineers and bioinformaticians who want to help us shape Lifebit — if this is something you’d be interested in, we’d love to hear from you: hello@lifebit.ai.</em></p><p>[1] <strong>Barkur S. Shastry, B. Jochimsen et al</strong>. <em>SNP alleles in human disease and evolution</em>. J Hum Genet. 2002;47(11):561–6.</p><p>[2] <strong>Bromberg Y. et al</strong>. <em>Neutral and weakly nonneutral sequence variants may define individuality</em><strong>.</strong> Proc Natl Acad Sci U S A. 2013 Aug 27; 110(35): 14255–14260.</p><p>[3]<strong>Mathers J et al</strong>. <em>The Biological Revolution: Understanding the Impact of SNPs on Diet-Cancer Interrelationships</em>.The Journal of Nutrition 2007; 253S–258S.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=4066e572eab8" width="1" height="1" alt=""><hr><p><a href="https://medium.com/lifebit-ai/deepvariant-as-a-nextflow-pipeline-on-deploit-4066e572eab8">DeepVariant as a Nextflow Pipeline on Deploit</a> was originally published in <a href="https://medium.com/lifebit-ai">Lifebit</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Improving genomic analyses over cloud using Grakn]]></title>
            <link>https://medium.com/lifebit-ai/improving-genomic-analyses-over-cloud-using-grakn-47cc8049d407?source=rss----34b6fcaa2544---4</link>
            <guid isPermaLink="false">https://medium.com/p/47cc8049d407</guid>
            <category><![CDATA[bioinformatics]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[genomics]]></category>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[orchestration]]></category>
            <dc:creator><![CDATA[Luisa Santus]]></dc:creator>
            <pubDate>Thu, 14 Jun 2018 13:37:07 GMT</pubDate>
            <atom:updated>2018-07-05T16:30:04.857Z</atom:updated>
            <content:encoded><![CDATA[<p>Why and how Lifebit is using Grakn to improve genomic analyses over cloud.</p><h3><strong>INTRODUCTION</strong></h3><p><em>Disclaimer: Long (but interesting!)</em></p><blockquote>These days, the words “genomics”, “genomic revolution” and “genomic data” are on everyone’s lips. <strong>Why is that? Why genomics and why now? What are the challenges of analysing genomic data?</strong></blockquote><p>In the last few decades, <strong>sequencing technologies have evolved incredibly fast </strong>which led to a <strong>drop in the cost of sequencing a genome</strong>. In 2003, as part of the <a href="https://www.genome.gov/12011238/an-overview-of-the-human-genome-project/">Human Genome Project</a>, the first human genome sequenced costed billions of dollars and the individual and manual effort of many researchers. Today, we reached a point where sequencing a <a href="https://www.nature.com/news/technology-the-1-000-genome-1.14901">genome costs only $1K </a>and only demands minimal human intervention. Soon, <a href="https://www.illumina.com/company/news-center/press-releases/press-release-details.html?newsid=2236383">this will drop to $100</a>.</p><p>The fast evolution in sequencing technology and the consequent drop in its cost translate into an <a href="http://journals.plos.org/plosbiology/article?id=10.1371/journal.pbio.1002195"><strong>unprecedented production of genomic data</strong></a>.</p><p>Not only is a massive amount of genomic data being produced, but this data is also extremely complex and we still do not fully understand the underlying biology and mechanisms. Hence, we are yet to see a set of standard go-to methods as these keep changing and evolving.</p><p>When someone, mostly bioinformaticians, try to make sense of genomic data, they usually follow a specific path to insights. This process is, in a lot cases, still very long and troublesome. At <a href="https://lifebit.ai/">Lifebit</a>, we identified some of its <strong>main challenges</strong>:</p><ol><li>Custom and inefficient solutions;</li><li>Time wasted on computational tasks; and</li><li>Limited pool of people who can actually produce these custom solutions for genomic data analysis.</li></ol><p>What’s more, when an in-house solutions can be engineered, sooner or later on-premise resources will become the bottleneck and computation/storage needs to be moved to the cloud. However,<strong>using cloud comes with three main concerns</strong>: technically challenging, security and costs.</p><p>At Lifebit we created a <strong>platform-as-a-service, called Deploit,</strong> which guides the user from data to insights and fully integrates with the way scientists already work. It advances their work through resource-management handling, cost optimization and learning from previous analyses — all this wrapped around an intuitive UX/UI.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*gLXnzOuqwVAmjQ7S.png" /></figure><h3><strong>WHY AND HOW WE ARE USING GRAKN</strong></h3><p>After such a long introduction, let us move on to the reason why we started this post: Why and how are we using Grakn to improve genomic analyses over cloud ?</p><p>Let’s focus on where Grakn makes a difference for us: <strong>what is the best resource to run specific analyses over cloud?</strong>We needed a technology that would allow us to both efficiently store our data and easily use it to train a ML model.</p><h4><strong>At Lifebit, when it comes to picking (or delivering) a technology, we follow three basic rules:</strong></h4><ol><li>The <strong>onboarding</strong> needs to be fast and a user should be able to use the basic functionalities of the technology quickly.</li><li>We care a lot about the technology’s <strong>learning curve</strong>: how long will it take a user to master the most complex functionalities?</li><li><strong>Evolvability</strong>: how stable and scalable is the technology over time as my needs evolve? In the case of a DB, we need it to not only model our data today but also be able to do so with an ever more complex and large dataset.</li></ol><h4>For us the winner ended up being……<strong>GRAKN</strong>!! :)</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/596/0*jrziq4GVItKBxVJg.png" /></figure><p>Once we identified the right technology, we had to define what kind of data should would be stored, and how it should be used, to find the best resources to run an analysis.</p><p>When a user comes to Deploit to perform genomic analyses over cloud, he first needs to define which kind of workflow he wants to use, then he picks the dataset to analyse and finally, on what kind of machine/resource he wants to run it on.</p><p>At Lifebit, we monitor <strong>real-time information</strong> of previous analyses, store it on <strong>Grakn</strong> and learn from them in order to provide suggestions to users on <strong>which instances are the most suitable for their analyses</strong> (depending on the nature of their workflows, data and needs).</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*bvtqgaxxnHdUjMRwCe1U9g.png" /></figure><p>Here’s a screenshot of a quick example of what Grakn looks like for us. Don’t be mad, for obvious reasons, I can’t share more than this :)</p><p>As mentioned above, the 3 main entities we use are: Pipeline, Dataset and Instance. Each has specific attributes; an instance, for example, can have a different number of CPUs, memory, storage, price, etc.</p><p>These 3 main entities are connected through the analysis relationship — ie. running a specific pipeline, on a specific dataset, on a specific machine.</p><p>After storing this information, we need to learn from it. We trained a ML model in Python which could give us the answer we wanted.</p><p>In this case, Grakn was again very helpful since it provides a Python client. Here is a small example of how easy it can be:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*ZjRUhDEv9c3n2Xpa.png" /></figure><p>This is basically how we are using Grakn at Lifebit: to provide suggestions on resource optimisation to run genomics analyses over cloud while minimizing costs and runtimes!</p><p>If you are interested in what we are doing and want to have early access to the platform please feel free to contact us at <a href="/service/mailto:hello@lifebit.ai">hello@lifebit.ai</a>.</p><p><strong>We are still at the beginning and we have big plans for the future! Don’t miss a thing and let us keep you updated on our following posts by signing up to our Newsletter or visit us on </strong><a href="https://twitter.com/Lifebit_Biotech"><strong>Twitter</strong></a></p><p><em>We’re also actively looking for great engineers and bioinformaticians who want to help us shape Lifebit — if this is something you’d be interested in, we’d love to hear from you: hello@lifebit.ai.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=47cc8049d407" width="1" height="1" alt=""><hr><p><a href="https://medium.com/lifebit-ai/improving-genomic-analyses-over-cloud-using-grakn-47cc8049d407">Improving genomic analyses over cloud using Grakn</a> was originally published in <a href="https://medium.com/lifebit-ai">Lifebit</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Donating your Genetic Code]]></title>
            <link>https://medium.com/lifebit-ai/donating-your-genetic-code-7384fc4cf3bf?source=rss----34b6fcaa2544---4</link>
            <guid isPermaLink="false">https://medium.com/p/7384fc4cf3bf</guid>
            <category><![CDATA[genomics]]></category>
            <category><![CDATA[dna-sequencing]]></category>
            <category><![CDATA[big-data]]></category>
            <dc:creator><![CDATA[Maria Chatzou]]></dc:creator>
            <pubDate>Thu, 26 Apr 2018 17:31:28 GMT</pubDate>
            <atom:updated>2018-06-14T12:50:58.910Z</atom:updated>
            <content:encoded><![CDATA[<p>The 12th of April was a very exciting day for Lifebit. As part of our commitment to transforming genomic data analysis, we are proud supporters of the GenoME mobile app, developed by the Personal Genome Project (PGP) UK.</p><blockquote>In fact, we sponsored the sequencing of one human genome — more about this below!</blockquote><p>Our London team had the privilege of attending the GenoME launch event in the magnificent Wilkins Building at University College London. It was a fantastic, memorable occasion. Uniquely for a scientific event, the results were presented by four Project Ambassadors who had donated their genome for sequencing, including Professor Stephan Beck, the Director of the Project. Professor Beck also gave an introductory overview, explaining the idea behind the project, the principle of democratising the analysis of the human genome and the enormous benefits we can derive from it as a shared resource.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*OWNCtfJvL4QGC9dhrlbbnA.jpeg" /></figure><p>The scientific part was followed by a music performance of an equally unique kind. The Benyounes Quartet played two pieces composed by Deirdre Gribbin. The composition was derived from the Ambassadors’ genome sequences by converting the four bases to notes using a transformation invented by Deirdre. Absolutely amazing. Deirdre gave a personal touch to the concert when she mentioned that her son’s Down’s syndrome inspired her interest in genomics, becoming a resident artist at the MRC LMB in Cambridge, where the DNA structure was discovered in 1953.</p><p>And now, to the (scientific) fun bit. The last part of the event was a raffle with a very special prize. Each participant was given a number coupon at the entrance, with five lucky winners to have their genome sequenced and published by the PGP. As mentioned at the beginning of the article, Lifebit is proud to cover the sequencing cost for one of the five genomes.</p><p>During the reception that followed we got the chance to chat (over delicious canapes and plenty of Prosecco) with participants and discuss various aspects of human genome sequencing and its countless implications.</p><p>The advances in algorithm development, hardware performance and networking have opened up tremendous new possibilities for bioinformatics. If genome research is to benefit from this progress we must ensure that scientists can focus on analysing the data and sharing the results instead of dealing with incompatible formats, proprietary setups and the like.</p><blockquote>__________</blockquote><p>At Lifebit we strive to facilitate the analysis of genomic data by removing the cumbersome and unnecessary layers of complexity that beset computational biology. Simplicity, reproducibility and openness are at the core of our technology development — a work driven by the scientists we have the privilege to collaborate with.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=7384fc4cf3bf" width="1" height="1" alt=""><hr><p><a href="https://medium.com/lifebit-ai/donating-your-genetic-code-7384fc4cf3bf">Donating your Genetic Code</a> was originally published in <a href="https://medium.com/lifebit-ai">Lifebit</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>