Downstream Analysis of Resistome data

To understand the downstream analysis of data, let’s imagine a story together. Of course, this specific case is fictional, but it could very well be real!

“Dr. Patricia just started her postdoc at a new institution after obtaining a PhD in molecular microbiology. In her postdoc project, she is interested in exploring the effect of antibiotics on the human oral and gut resistomes. She started by conducting a pilot study with, let’s say, antibiotic X. She recruited 10 patients that had received the drug and 10 patients that were exposed to a placebo control. From these patients, fecal and saliva samples were collected. As we have seen in the previous week of the course, samples were properly stored, and DNA was extracted according to protocols. Libraries were prepared carefully and sent for sequencing. Once finished, Patricia received the FASTQ data, and she quality trimmed and filtered all the files. Further, she filtered out the human DNA contamination using BowTie2 software in order to obtain high-quality reads. Next, she used the AMR++ pipeline and MEGARes database for identification and quantification of to identify and quantify ARGs in all fecal samples. As a result of this process, she finally obtained a table containing abundance information (read counts) of hundreds or thousands of identified ARGs across all samples. Now, she wants to make sense of and test some of her own hypotheses with the experimental data obtained. For example, she asks herself:

Are there more antibiotic resistance genes (ARGs) in patients treated with the antibiotic X compared to controls?

Is there an enrichment of some specific ARGs or class of ARGs in individuals treated with antibiotics?

Does the resistome gets affected by the bacterial profile OR Are there any associations between ARGs and the microbes present in the sample?

Apart from these specific questions, she also wanted to explore the data to gain some more meaningful valuable and biological insights, without a specific hypothesis beforehand.”

With this story in mind, we can use various visualisation and analysis approaches in order to understand and interpret the information present in multidimensional count abundance tables, which is also known as Downstream analysis of resistome data. Such analysis can be generally divided into four different categories:

Composition profiling: to visualise the composition and characterise the resistome using alpha and beta diversity approaches developed in community ecology such as alpha and beta diversity.

Comparative or Statistical analysis: to identify features, such as antimicrobial resistance genes, that are differentially abundant between conditions, under investigation. For example, Treated Vs. Control (Patricia’s example).

Integrative analysis: to explore the potential associations between the microbiome and resistome profiles using several omics integration approaches,. For example, understanding whether the bacterial composition shaped the gut resistome in the sample. whether the gut resistome was shaped by the bacterial composition in the sample.

Functional profiling: to analyse resistome at various functional categories or levels, such as drug class or mechanisms, in order to explore biological and phenotypical insights with regards to the data.

We will go through each of these types of analyses in detail in the coming steps of the course. But before going further, let’s talk a bit about various challenges associated with such type of data and its analysis:

Data is complex (multidimensional, too many zeros (sparse), uneven library size, skewed distribution, compositional, etc., causing several statistical challenges).

Not everyone knows computer programming or have Bioinformatics skills (E.g., Clinician and Bench researchers) to analyse such data.

Lack of standards or one best method (Analysis is exploratory in nature).

So now the real question arises how can we can tackle these challenges?

One solution to all problems would be to have a tool or software that is very easy-to-use, freely available, with a simple and interactive interface that would enable us to perform various complex analyses and visualisations in an intuitive manner intuitively.

In this week of the course, you will learn and perform downstream analysis and exploration of resistome data using the tool called ResistoXplorer.

In the comments below, please let us know:

• Are you currently running or involved in a resistome study?
• About the resistome studies that you are doing or planning to do in brief.
• Which bioinformatics pipeline/software have you used to perform upstream analysis of data?
• Are there any general or specific hypothesis that you are interested in testing, or do you plan to explore the data to generate one?