 Good morning, my name is Maria Yambriq and I will introduce you Amanida, honor package for overall results metanalysis. Let's start with a brief introduction. So metanalysis is the statistical combination in a single estimate for results from a study answering the same question. This is used in medical research, and before a clinical application, we will need a lot of academic findings to be able to do the metanalysis. These academic findings come from different files, for example, genomics and others, and there is one that is now emerging as a clinical tool, which is the metabolomics. Metabolomics is a study of endogenous and exogenous metabolites in biological system. The aim of this omics is provide information about all the metabolites that are in a system. These metabolites are small chemical compounds. For this, the diversity is really huge. For example, in the human organism, we have identified until now more than 200,000 metabolites, but this number is increasing day by day, which is the common workflow for metabolomics. This starts with a sample, can be a fluid, can be a cell. And then this is analyzed through high-triple technologies like mass spectrometry or nuclear magnetic resonance. And we obtain a lot of results. From one sample, we can obtain more than 100 compounds identified, and a lot more than we cannot be able to identify. What we do with this list of compounds from each sample, we do some statistics. This is the statistics, more common use. So first, what we do is compare. We compare two groups. We will have the control group versus the disease group. For example, cancer. In this case, the first thing is to know the statistical significance. We want to know if one group is different or not from the other. Once we have this statistical significance, we will know the biological significance of who many change there is between the two groups. This is known as the full change. And what we have is a ratio between the population and the control. These are the statistical that are used in metabolomics. So when we want to do a meta-analysis with the metabolomic results, we find some problems. In the case of the traditional meta-analysis, we will need some metrics to be able to perform this. For the standard effect size, we will need the mean that we can find on the metabol... and the metabolic studies. But we will need the standard deviation, or standard error, that this is not disclosed in the metabolic studies. We don't have a standard initiative for statistical reporting of the data, and this is why we cannot find these estimates on the articles. And the number of participants that this is disclosed in the studies. The other way is to perform the meta-analysis with metadata. The amount of data from metabolomic studies is really huge, and this is really cost, so it's not an option really common to perform. For this, we have developed the Amanida meta-analysis approach. This meta-analysis combines only two metrics, the p-value and the full change. So first, the p-value is combined using the Fisher's methods, weighted by the sample size. So it's not the same that we have a p-value obtained in a bigger study of more than 200 participants, than a p-value obtained from a small study of, for example, 20 participants. For this, we combine also the full change, which is logarithmically averaged, and it's also weighted by the sample size. This rationale is implemented in a grand package called Amanida. With this package, you can use simple text files and do the meta-analysis approach. We have also included another type of meta-analysis, more qualitative, then we will explain deeper later that this is the boot content. Then we have this graphical visualizations of the results and also of the data. Here is the type of data that Amanida needs to work, so we will need the idea of each combo. That can be the common name, it can be a number, whatever, the user one, then obviously the p-value, the full change, and the sample size. So with only these four columns, we can perform the meta-analysis. Then we obtain the results, which is a table with the combined p-value and the combined full change. If we're talking about metabolites, Amanida also allows to complement the information about these chemical compounds. So you will have the molecular formula, molecular weight, and also the idea from public metabolomic databases. Here we have the visualization of the results for Amanida. In this case it's a volcano plot, like we have in genomics, so the full change logarithmically transformed is plotted versus the logarithmically transformed p-value. And we have level the components that are significant, either for up-regulation or the regulation. Then we move to the qualitative part. So there are some cases that the full change is not numerically disclosed, so we only have reported if the component is up-regulated or down-regulated. So in this case, what we do is about containing the same about to each component depending on the trend. If the component is up-regulated, we will assign a plus one. If the component is down-regulated, then it's assigned a minus one, and the sum of the words will be the total result. In this case, for example, citric acid, which is reported in six articles, which analyzed urine for colorectal cancer, the total sum of the votes is minus two. Then as before, you can complement all the information from the public databases and visualize the results. At right, we have the components that are up-regulated, at down, at left, we see the components that are down-regulated. So what happens with these components that are not troubles between the studies? It happens, usually in metabolomics, that one component is reported up-regulated in one study and down-regulated in other studies. So we developed the exploratory plot to see at first glance which components are these that are reported in both sites. For example, as before citric acid is reported, we say that it was reported in six articles, and here we see that there are two articles that say that it was up-regulated and three articles that say that was up-regulated, and one article that say that have no change between controls and patients. All these have also been implemented in a shiny app to be more user-friendly. And in this case, it works similarly as the grant package, so you can upload your data, you choose if you want the quantitative or the quantitative analysis, and then select the columns that you want to use. We have to remember that if we have missing data, it will not be taken into account for the analysis. We select our data, we calculate them, we show here the rows that are not taken into account, and then we have the results. Here we have the work on a plot where the user can choose the cutoff for the p-value, a statistic of significance, and the cutoff for the full change, the biological variance. And only with a click, you will know the component that is significant for up-regulation and down-regulation. The same of the packets, you can see the complete table of the results of the Amanida meta-analysis approach. We include also the vote counting and the explorer plot to see differences at first glance. At last, all the results can be downloaded directly in an active metal file where you have all the tables obtained with the meta-analysis using the Amanida approach, and also the vote counting including all the graphs that can be done with the Amanida error package. So, this is all. I want to thank my supervisors and the people that helped me. And thanks to you for hearing me. If anyone have any question, I will be glad to answer it.