Metrix is an application written in java which will run in the background of the server much like a daemon.

To answer one of the first questions of Brett, the code does not necessarily gets executed at one single point in time. As Metrix is running continuously in the background it monitors the illumina run folder for any changes (New folders / files being created).

If in this run folder a new folder is created which matches a certain pattern (YYMMDD_sequencername_runnumber_flowcellbarcode_OptionalRunName) then it is added to the list of folders that Metrix needs to monitor. Several validity checks are run to ensure that the folder created is indeed a valid illumina run folder.

In the first pass of finding a valid illumina run folder, Metrix will check how much information / data is readily available to parse. All the parsed data of a run will be stored in a custom made java object called a "Summary" and are in turn stored in an SQL database.

I have ensured that no modifications to the file system take place, in other words, Metrix will not write any changes to files present in the Illumina run folder or anywhere else on the file system. The only data being generated and stored is by Metrix in the MySQL database.

Similarly as long as the Metrix daemon is running, it can be accessed by a piece of code called metrixClient. This code allows the interrogation of data and information known to Metrix (stored in the database) by querying with parameters. I have made Metrix in such a way that it can generate the output in multiple formats (XML, TSV, Java object), making it suitable for GNomEx as well as CLI (Command Line Interface) purposes.

Next I’ll explain a bit what I envisioned of how Metrix and GNomEx are supposed to work together. The raw data generated by the HiSeq's and MiSeqs hold an enormous amount of valuable information. Typically this is all the information you can see in Illumina's well-known Sequence Analysis Viewer, but because we can extract all the data ourselves using Metrix we can create our own graphs and generate reports on various metrics such as Machine efficiency, Cycling times, Q20/Q30 %’s, Phasing/Prephasing, Cluster Densities and Indices.

As Brian already said, only upon flowcell assembly completion the folder name is known in GNomEx. Because as it stands now Gnomex and Metrix are separated, I propose that make a specific component in GNomEx where admins can link assembled flowcells with the actual run directory via Metrix. This can be done by passing on the used flowcell ID to Metrix which in turn can return the full path of where the data is stored, along with any requests sequencing metrics that are required.

For now I have chosen to restrict the Sequencing Progress view on the dashboard to admin users only. Usually when a run is ready we get notified by our standalone version of Metrix and we start the demultiplexing process. The demultiplexing takes a while as well and once it is done we will notify the clients.

In the long run my desire is to create reporting framework within GNomex where graphs and tables can be created using Experiment data from GNomEx together with metrics data parsed by Metrix.

Regarding information shown on the dashboard. This is fully customizable from the side of GNomEx as well as the side of Metrix. We could condense the initial view on the dashboard to merely a progress bar and the run name and provide a button which will show you a popup with more detailed information.

Contact Us

Research Informatics Director
Andrew Post, MD, PhD

Research Informatics Associate Director
Shirleen Hewitt, DBA


HCI Senior Director Oversight
Aik Choon Tan, PhD

Faculty Advisory Committee Chair
Aik Choon Tan, PhD

Faculty Advisory Committee Members
HCI Research Executive Committee