New in MetaGenyo? Go to Help tab.
You can try these example datasets from published papers:
If you use MetaGenyo, please include this reference:
Martorell-Marugan J, Toro-Dominguez D, Alarcon-Riquelme ME, Carmona-Saez P. (2017) MetaGenyo: A web tool for meta-analysis of genetic association studies. BMC Bioinformatics. 18:563
model: weighted regression with multiplicative dispersion
predictor: standard error
MetaGenyo is a simple, ready-to-use software which has been designed to perform meta-analysis of genetic association studies.
MetaGenyo requires a specific data format to worth with. The first n columns of your file can contain any type of data of the different studies that will be used at the Subgroup analysis step. N is defined by yourself. The following columns must be exactly like this:
The meaning of each column are the following:
In conclusion, your data should be something like this (remember that previous columns to "AA cases" column are chosen by yourself):
You can write the column names that you want to (but there MUST be column names). But there are some exceptions: column names with genotyping data must start with the genotype (AA.cases it's correct, but cases.AA it's not). Missing values are not allowed. Be sure that your data is in the correct order.
Once you have all your data, you must save it in a suitable format for MetaGenyo. The tool accepts Excel files (.xls and .xlsx) and plain text files.
Once you have prepared your data, you are ready to use MetaGenyo. First, enter to the webpage, press "Browse..." button and select your data file. Now:
Once you have submitted your data, you must click "Your data" tab to review it. If your data looks nice and no error appears, it means that you submitted your data correctly and you can continue with the following steps. Otherwise, you must return to "Data input" tab and review that options selected correspond to your data characteristics. Maybe you should review your input file too.
In this tab, you will find your data with two columns added: HW.P.value and HW.adjusted.P.value . The first one contains the p-value of chi-square test for Hardy-Weinberg Equilibrium (HWE) for the control data of each study you included. The second one contains these p-values corrected for multiple testing by FDR mehod. This is a way to check study quality: if p-value is significant, control genotypes may not be in HWE and you should be careful with that study data or exclude it. That's because departures from HWE can occur due to genotyping errors, selection bias in the choosing of controls and stratification . If you want to exclude some study, just delete it from your data and resubmit your file.
You can download the results in Excel (.xls), comma-separated values (.csv) and tab-separated values (.tsv) formats. Simply choose your desired format and press "Download" button. You can also interact with the results in the table showed.
In this tab, you will find association test results calculated for the comparison that you choose on the left panel. These results are:
In this tab, you will find a forest plot for the comparison that you choose on the left. A forest plot gives a visual representation of the variation between the studies included in your meta-analysis for the selected comparison . It also provides an estimation of the global result of all the studies.
In this tab, you will first find a funnel plot for the chosen comparison on the left. This plot is a graphical test to check if publication bias exists in your meta-analysis . If the funnel plot is similar to a symmetrical inverted funnel, there is not publication bias in your data. On the other hand, if there is asymmetry maybe publication bias exists ini your data, so the results of your meta-analysis must be interpreted carefully .
In this tab you will also find Egger's test results for your data and selected comparison. This is a more objective way to detect publication bias . If Egger's regression test p-value is lower than 0.05, publication bias is possibly present in your meta-analysis.
In this tab, you can choose a column of your data and stratify the studies based on this column. A table will be generated with some results of the meta-analysis for each of the groups formed. The test of association is calculated with fixed or random effect model, depending on the test of heterogeneity: If p-value < 0.1, random effect model will be used. Otherwise, fixed effect model will be used instead. If there is any group formed by one single study, this group will not appear in the results (obviously, meta-analysis can't be performed with one study).
These tables can be very useful to easily detect statistically significant associations in subgroups when no results are obtained including all the studies. The generated table can be also useful as a summary if subgrouping column is not selected.
In this tab, you will find a forest plot for the robustness analysis by leave.one-out method. The analysis is repeated excluding one study each time in order to visualize if any study has a significantly greater contribution to overall statistics that the other studies.
If you have any doubt, question or suggestion, you can write us to firstname.lastname@example.org .