BLT Stages

The microbiology group at SCKCEN ran a first of its kind microbial growth experiment in the international space station. Four specially designed bioreactors were send to the ISS to compare the growth of cyanobacterium Arthrospira sp PCC 8005 in space with the growth on ground in the lab. Next to proteomics, DNA methylation and several biochemical analyses several growth parameters and data about bioreactor performance were collected.

Four consecutive batches were run on each reactor spanning in total five weeks of runtime. To compare different batches and different reactors and to integrate them with, for example, the proteomics data, a tool for clear visualization and options to select different batches and plotting options is needed. The data from more than 35 growth- and bioreactor-performance-related variables consist of 3 million data points for each reactor, and is stored in multiple CSV files. These files are too large to visualize and edit with graphical interfaces such as Microsoft Excel.

Furthermore, the data is recorded every second and stored in a log-file, regardless of the state of the reactor (active or idle). For example, optical density measurements are done every 12 hours, so only once per 12 hours, the recorded value changes. For proper and informative data use, the exact time of measurement with the corresponding value should be used. Extracting these points is challenging as the logging does not indicate where the recorded values change.

Another difficulty is that the experiment did not ran completely as planned, i.e., the running schedule of the experiment had to be adapted during the experiment. This also needed to be taken into account whilst organizing and extracting the data. 

In addition, for the flight reactors, the provided data is not divided in the different batches. Therefore, the batch exchanges points need to be searched in the log-files. Although the batches were planned on fixed dates, the exact timing relied on the astronaut timing and adaption of the running scheme made based on observations during runtime. 

R scripts were used to extract, organize and visualize the data. First, an R script was written to combine the reactor-specific data, search for the batch exchange points and extract the relevant time points. Next, a script was written to adapt the data-format for easier use in a Shiny dashboard. Finally, the R package Shiny dashboard was used to create a user friendly dashboard to visualize different bioreactor parameters and choose which reactor(s) and batch(es) are plotted and allows the user to download a table with selected data points.

The Shiny dashboard allows the user to visualize six parameters separately. It is possible to select data from one or more batches and one or more reactors. The generated plots allow zooming, saving, setting a grid to clearly separate batches and toggle between different x-axis input. It is also possible to lay the different batches over each other and a clear difference in marker shape and color is set for each batch and reactor. The selected data can also be viewed and downloaded as a table.

There is still ongoing work on the dashboard to add some additional parameters that need more calculation, make it possible to visualize text written errors or running scheme adaptions on the plot at the time of occurrence. A CSS-styling with the company colors could also still be added.

The developed dashboard will also be used for the follow up experiment where real-time follow up of the bioreactors is needed. Discussions are ongoing on how the data coming from the ISS will be made available for real time plotting and will require a script to fetch the data automatically

 

The ever-growing demand for data is bigger than ever. In a continuously evolving world of computing, methods have to be developed to make this big amount of data interpretable and understandable. In the field of microbiology, metagenomics is becoming a key player in determining bacterial, archaeal and even eukaryotic species present in vast and complex microbiomes. Culturing microbiome strains has proven to be challenging, as has the process of determining species from shotgun sequencing approaches. After sequencing, reads are assembled into contigs, which then serve as basis for scaffolds, which can finally be used to close genomes. The clustering of contigs (also known as binning) is a crucial step in correctly assembling a genome from metagenomics samples. The performed research, attempts to identify suitable binning programs based on the binning output generated. The selection of these programs is based on a comprehensive literature study, identifying CONCOCT, MaxBin2.0 and MetaBAT2 to be most suitable and user friendly for the tested sample of interest. A python script was developed to support analysis of the output of metagenomics tools. This script generates a summary per selected binning program, yielding information about its overall performance, as well as provide summary statistics for the sample of interest. 

 

By detecting marker genes like the 16S rRNA in bacteria, an overview of the community and its relations between species can be studied using high-throughput sequencing technologies. To grasp a better understanding of 16S rRNA metagenetics, several pipelines (like mothur, QIIME, OCToPUS or LotuS) have been designed to process these sequencing results and allows us to investigate these microbial communities. An in-house pipeline has been developed, within SCK●CEN, that performs post-processing analysis of the output of these pipelines. Yet there was a need to visualize these results and present it for the users in a comprehensive report. Here we present a visualization and analysis pipeline developed in R. This tool takes the output of the SCK●CEN in-house pipeline, creates a report describing the alpha and beta diversity, performs biomarker detection and presents hypothesis testing based on the meta information of the experiment. As a demonstration, a mouse experiment was processed using the OCToPUS pipeline, followed by an analysis and visualization using the post-processing tool. The tool will be later integrated with the OCToPUS pipeline into one comprehensive package for assessing any 16S metagenetics data.