Open mfazel opened 2 years ago
Thanks for the question! We can definitely look at improving the top loci table in the future.
In the interim, you may find it convenient to know that the manhattan plot shows much of the same information as the table, and that you an also navigate to loci by clicking directly on significant points in the manhattan plot.
Looking back at this proposal in more detail: is it correct to say that you want a way to find a particular variant, and that you want to browse all hits in sequence?
(eg, you want to go to row 171 because it is the next hit- the position in the table isn't otherwise important as long as you can find an entity of interest)
I wonder if two navigation features in the "region" page would help solve your problem better?
chr:pos
specifier to jump to a specific region of interest.Can you offer an example of where searching the table would be preferable to looking for your locus in either the manhattan plot, or via the region page?
Ok, the question is, say there are number of known peaks for a trait (can be in a gene or in a farther region), then we want to know if there are any new peaks close or slightly far from those regions in our data and we don't know them completely. I used the batch view and it's good, but user has to know the coordinates and type them one by one or go over all peaks one by one to find out if there are any new ones. As I said before, user can sort the table by Marker and then go through them, the problem was that every time, the table gets back to original order. Using the Manhattan plot also was a little bit tricky, when there are multiple peaks very close to each other. The closest feature I could use was the search box in the Region View after clicking on one of the peaks.
Here I simplify it, maybe I did not know how to do it. How do one gets all significant peaks, say on chromosome 18, in the table format. It would be good if the table had chromosome and positions in separate columns for sorting purposes. Also, making the number of visible rows in the table adjustable would be useful too. Thanks, Mehdi
I see. By default batch view sorts "top hits first".
You'd like to see that combined with a filter based around the range of positions? We've been tossing around some concepts for this sort of thing internally; I'll give some thought to this, though we might have to pilot the concept in some other pending work first.
Action items:
Yes, something like that, but again it doesn't need to be fancy and I think a couple of features of java script table views are very helpful and you see them almost in every website. (I'm guessing this is also done in java script!) Here is an example link (and just in case I include the screenshot too). You see there is an adjustable number of rows by user (10,25,50.....) on top left, a general search on top right and download link below the table and user can search for anything not only coordinates. Some other websites also include searching on each column header separately to allow for "AND" operation. Thanks,
Hi,
I use locusZoom occasionally to brows some GWAS data and create plots, which I'd say, they are nicely generated compare to amount of time a user spends. This is just a suggestion not a big problem. Looking at the Loci table right bellow the Manhattan plot in the study page. By default it's sorted by Pvalue, which is fine, (and I assume the table presentation is done in java script) but there is no search option for the table, not a general search nor a column by column search box. The number of rows in each page is also fixed on 10 and it's not possible to extend the table to more rows to see more rows easily. The problem is if a study has 20 different pages (10 row each), and user wants to go to row number 171, has to go through all pages until that row and then if I go to that locus and then come back, I'll land on the original table sorted by Pvalue (unless open it in a new tab).
Thanks for your nice tool.