[jaempawi]

Summary

Fixes from two reviews of the catalog as a visitor. Layout, links and wording, plus corrected values (listed under Corrected values) and cis-QTL coverage taken from Synapse. No dataset, ID, status or access value changes.

Layout and links

• Centered pages. On wide screens the pages sat against the left edge. The text column is now centered next to the side menu. Home is centered at up to 1,200 px; catalog, inventory and coverage pages at up to 1,480 px.
• Site-name link. "FunGen-xQTL Resources" in the header linked to the server root (/), which left the catalog on adsp-fgc.niagads.org. It now opens the catalog home.
• "On this page". The list no longer shows "Related datasets" on the seven GWAS and reference pages that have none.
• Explorer links. The README and the resource description pointed to the old wanggroup.org address. Both now use https://xqtl.niagads.org/AD-loci-explorer/.

Wording

• Data access. The README said all datasets are controlled access. It now says the Synapse folders are controlled and the NG00184.v1 summary statistics on NIAGADS DSS are open.
• Home: "AD GWAS studies". The GWAS list is renamed, with a note that the four studies give the eight GWAS datasets on the AD GWAS overview page.
• Access text. Cohort and GWAS pages said "Access terms for this page are not set yet" while the table below showed Controlled. When a page has a Synapse folder and no access value, the text now matches the table.
• Study pages (ROSMAP, MSBB, Knight ADRC, MiGA, STARNET). Empty metadata lines ("NR", "nan") are removed, and "Website&Logo" reads "Website".
• GWAS pages (Bellenguez, Wightman, Kunkle, Jansen). The "Path(s) to summary statistics" lists are replaced by a "Source files" paragraph. A note naming who approved an internal data request now states only when the data were obtained.
• ROSMAP DLPFC eQTL page. The "need to replace" placeholder link and five empty Dataset Details lines are removed, and two typos fixed.

Inventory and coverage

• In hand vs Staged. "In hand" was described as "usable now" while every dataset page says Staged (uploaded, not yet shared). It now reads "Obtained and on Synapse or DSS. Release status is shown on each dataset page."
• Inventory counts. The All and Next phase counts leave out the "Other new or expanded datasets" placeholder row (All 37, Next phase 1).
• Coverage note. The note says which datasets the grid includes and that a dash means no status is recorded, not that a method does not apply. The dash tooltip says the same.
• Wide tables. The coverage grid and inventory table are named regions that can be focused and scrolled with the keyboard.

Coverage

• cis-QTL coverage. The grid showed no cis-QTL status for any dataset. It is now marked done for the 20 datasets whose cis top-loci file from single-context fine-mapping is on Synapse (syn69865687). The note under the grid says so.

Catalog on phones and search

• Sort on phones. The phone layout has no column headers, so it had no way to sort. A Sort menu (dataset, cohort, type, N) now sits next to Filters and keeps the sort in the web address.
• Cards show N and location. Phone cards now show the sample size and the Synapse location.
• Search results. Dataset results in the header search show the page's full title, adding the modality when the title lacks it, so the eight "ROSMAP DLPFC" results can be told apart.

Corrected values

• Bellenguez 2026. The page guessed what noproxy, main and nobiobank contain from the file names. It now cites the source study (Bellenguez et al. 2026, Nature Genetics, doi:10.1038/s41588-026-02583-1), describes each analysis from the paper, gives the main analysis sample sizes, and names Haochen and Yining as running the fine-mapping.
• LD reference panel. The sample size read 16,905 in some places and 16,571 in others. It now reads 16,571 everywhere, matching the panel description.
• PsychAD. The inventory note gives the donor counts by brain bank as listed on Synapse: 1,494 donors (1,042 MSBB, 300 HBCC, 152 ROSMAP). It used to say "1,042 MSSM" and left out HBCC.
• MAF cutoff. Five QTL pages gave the minor allele frequency cutoff as 5/(2.0592) = 0.00422297, one example denominator copied across datasets with different sample sizes (one page also showed 5/(2.0416) = 0.00422297, which is 0.0060). The TensorQTL workflow computes the cutoff as 5/(2N), with N read from each analysis's covariate file. The pages now state that rule.

Left for the data owners

• Every dataset is still "Staged", as confirmed.
• cis-QTL is left blank for seven datasets with no cis top-loci file on Synapse that clearly matches them: Knight, MSBB, ROSMAP DLPFC, PCC and single-nucleus sQTL, the two metabolomics QTL datasets, and ROSMAP microglia eQTL.

Checks

• Local Hugo build. The 32 browser checks pass, including no sideways scrolling at 1440, 1280, 768, 390 and 320 px. The internal link check finds no broken targets.
• Phone (390 px): Sort menu reorders cards and updates the address; a ?sort= address sets the menu; the menu is hidden on desktop.
• Coverage grid: 20 cis-QTL cells marked done.