I'm afraid that you lose. Recall that a table is a set, that is, an unordered object. So what you have are just some dotted random unrelated annotations. There are two ankles, there are two migration percentages and some more junk. But there is nothing that says which migration percentages is related to which ankle. If there is any relation to any ankle at all, that is.
You need to go back to the data source and make sure that data is stored in a manageable way that upholds any relations that might have been there initially.