 Hello and welcome to SIB in CircoTalk on the MHC Motif Atlas. I'm Daniel Tadros, a PhD student at David Feller Group of Computational Cancer Biology Lab, and we will be introducing you to a new database of MHC bandic specificities and likings. It is based on the recent publication and website developed by our lab, and before starting the talk, I would like to thank Simon, who is the main architect of the website for his contribution in this project. The target audience for this talk are researchers interested in immunology, cancer immunotherapy, vaccine development, and machine learning. First, I would like to introduce you to the main step of the interaction between infected and T-cells. The T-cell activation is governed by two main steps, the antigen presentation pathway, where the protein gets degraded by the proteasome into peptide. This peptide gets then transported into the ER via TAP, which some of them bind to the MRC molecules, which is the most selective step in this process. And then the peptide MRC complex is presented at the cell surface for GCR recognition. That's why MRC likings are considered as promising therapeutic targets that have been widely used in preclinical and clinical studies. For instance, in cancer immunotherapy, MRC likings have been used as personalized vaccine to boost the immune system to recognize neo-antigen. In addition, studying the presentation pathway of these likings help us to better understand the disease and the clinical process required to fight against it. There are two clusters of MRC molecules. The MRC class one are recognized by CD8 T-cells, bind short peptides roughly between 8 to 14 residues, and the likings primary anchors are at the second and last positions. And for the class two, they are recognized by CD4 T-cells, bind longer and wider range of peptides roughly between 12 and 25 residues. However, the likings have a fixed binding core of nine amino acids. In addition to the different characteristics between MRC class one and two, MRC one and two genes show a very high degree of polymorphism, and thousands of different alleles have been documented. However, only a few hundred alleles with available likings. As a result of this polymorphism, different alleles show different binding specificities and bind different repertoires of likings. So many researchers capitalize on large MRC data sets to train machine learning tools to predict MRC likings that could be recognized by T-cells. Over the last decade, vast spectrometry-based MRC likings data have become the dominant source of information about MRC binding specificities. These data enable researchers to determine binding motives for hundreds of MRC alleles. Here's an example of a binding motive of an allele. This is a list of likings of same lengths. We can clearly see the binding specificity of this allele at position two and nine. However, MRC liking prediction tools are often used as black boxes and do not necessarily provide explanation on why a peptide gets a good or a bad score. So to facilitate the understanding of the main binding properties of MRC molecules, we present here the MRC motive atlas or an intuitive visualization of MRC binding properties and a way to interpret the MRC liking predictors' results. Here's an example of the binding specificity of an MRC class one allele. We provide a binding motive for each peptide length separately as we can see there is a high similarity in the binding anchors between different lengths. We also provide the peptide length distribution of the allele and we can see there's a high preference of peptide of length nine. Also when it's available, we provide different binding mode of the allele. We can see for example the arginine can only be in position three or six but not in both positions at the same time. And finally the motive of the phosphorylitic ligands when it's available presented in purple. Similarly for MRC class two as previously mentioned MRC two alleles bind to long peptide but with fixed binding core of nine acid. We also provide the petal length distribution for the allele and we can see there's a high preference of peptide with length 15. And when it's available provide different binding mode for the allele we can see for example the arginine and lysine can only bind in position four or six but not in both positions at the same time as it's shown in the first motion. Here is a summary of what we provide on the MRC motive atlas. We collected more than 500 samples from different species that resulted into more than one million ligands interacting with few hundreds of MRC one and two molecules. We provide thousands of predicted MRC binding motives based on our MRC motive predictors of class one and class two. We also provide direct links to MRC x-ray structures and list of MRC sequence available in our database. So now let's have a look of the web interface of our motive atlas. Here's the home page for the MRC motive atlas. We have four main pages the class MRC class one, class two, tools for different predictions and the frequently asked questions page. So let's explore class one. As you can see here you can explore different binding motive for different lengths for example also for multiple specificities, phosphorylated ligands and you can download a list of the unmodified ligands or phosphorylated ligands for this specific allele. You can also check direct links for the PDB structures for this allele with different ligands. What's also interesting in this page that you can only compare different alleles at the same time which is really interesting in our exploration of the binding specificity for different alleles. And here you can also download all the available data for class one from unmodified to phosphorylated ligands both of them MRC sequence and the x-ray structures. Now let's go for class two alleles. Similarly as class one you can see the multiple specificities. You can also compare between different alleles at the same at the same time. But in addition we provide an extra page for common properties between class two alleles that can help you to better understand the binding specificity for class two alleles. As for the tools page we provide direct links for different predictors for example MRC lightning predictor for class one and class two, image elasticity predictors, MRC motif deconclusion tools and the direct links for the MRC lagging databases. So if you have more clarification you can also check the frequency ask question. With that I want to conclude this talk by first thanking David for this exciting project and Cimo and Julien for all their effort and contribution in this project and you for taking the time to watch this video.