MalAvi
Avian haemosporidian database

MalAvi · for curators

A guide to
curating MalAvi

When data are submitted through the standard template, you will get an email with a pdf report. You read the report and evaluate the submission and then you click a link in the pdf that takes you to a Google form (autofilled with the submission info) where you record your decision ("Accept", "Flag", or "Reject"; the latter two require written explanations).

If someone deposits a PDF in the system without a data template filled out, you will still receive an email but no PDF report. The email will prompt you to extract the data yourself, fill in a submission template, and submit it yourself. I have a program that can try to extract data from the paper, but it doesn't work for most papers because host-parasite associations are often not reported as tables, or there is extra info that is hard to capture for some other reason.

A diagram of the process

A data template file submitted by a user Submission is checked and a report generated Curator receives the report for review by email Curator clicks the link in the report and records a decision A PDF submitted with no data template Curator receives an email, but no report Curator reads the paper and fills in a data template submitted like any other

The emailed report

Live

When someone submits the data template file, a report is generated and it comes to you as an email from “malaviadmin@gmail.com”. The report goes through several dozen automatic checks of the novel sequences, the proposed names of new lineages, and the host and geographic data. But these are just some basic checks, you still need to review.

In case someone tries to reserve a lineage name that is wrong for some reason (e.g., already taken), the report will propose the next available name. If you accept the submission, you are also approving the suggested name (unless you flag it with a comment).

When you click the button in the report to make a decision, it will link to a Google forms doc that has all the submission info pre-filled out, and you will just have to answer a few questions.

Add malaviadmin@gmail.com to your contacts because that's the account sending the emails. The email is an ordinary Google account and it seems to be delayed getting to my university address, but it does work.

The MalAvi notification email: who submitted, their institution and country, whether the data is published, what they are sending, how many files, their notes, a statement that the automated checks have not run yet, and a link to record a verdict.
A real notification, from 2026-08-06 (testing phase). The link it shows looks strange because my university's email service rewrites links, presumably for security reasons.

Reading the report

Live

The report is a PDF that opens in the browser. Here is an example report built from fake data to explore.

Example curator reportA full report built from fake data, so you can see what you will be sent before you are sent one.PDF
Four checks from a real report. Two are marked blocking: a proposed name that MalAvi already has, and a sequence identical to an existing lineage. One is a warning about sequence quality. One is information about host and geography.

Recording your decision

Live

Click the link in the report and it takes you to a short form to fill out. Should be easy.

The verdict block at the top of a report: the submission identifier and revision number, and a button reading Record your verdict on this submission.

The three options you will be choosing:

Accept

This can go into the next release, as-is.

Waits three days before anything happens, so another curator can also evaluate, if necessary.

Flag for further review

Choose this option if some part of the submission needs work, or you are unsure about something and want another curator to review as well. A flag is the right answer for anything that could be corrected: a duplicate name, a chromatogram to re-read, a wrong host or locality.

Blocks the submission until the curators can resolve it. The submission keeps its number and its reserved lineage names, and somebody (usually the lead curator, with you in copy) emails the submitter what to change.

Reject

This should not enter MalAvi at all. Nothing is sent to the submitter automatically, and when a lead later closes the submission, the notice carries no reason and the reserved names are given back to the pool.

Blocks the submission the same way a flag does. See below. The lead curator can still overrule the decision after discussion with the other curators.

A rejection blocks a submission, but does not finish it. When the curators have come to a final decision, the lead curator chooses “Close a submission for good” on the same verdict form and records a reason. That step ends the submission, meaning that the reserved lineage name(s) become available again, the submission leaves the public queue, and the submitter is informed three days later; the delay is in place to give curators time to catch and fix a mistake in case one was made.

A submission somebody has already accepted cannot be closed. Flag it first, so the objection is on the record with your name against it.

Why flag rather than reject when something is fixable. Closing a submission gives its reserved lineage names back to the pool. A submitter who spends months re-sequencing because we asked them to would come back behind anyone who claimed the name in the meantime. A flagged submission keeps its names and its number while the fix is made; the corrected workbook simply replaces the old one and a new report goes out.

When something is wrong with a submission

Live

If you want to make a correction, you must also flag the submission. In other words, a correction can't be made AND have the submission accepted at the same time.

Here's the breakdown of what a correction would look like:

  1. Flag the submission, and explain what's wrong and how it should be fixed.
  2. The lead curator has to approve after discussing with you and potentially the author who submitted the data. No correction is applied without that approval, and a lead cannot approve their own.
  3. The maintainer makes the agreed upon change.
  4. The corrected submission gets sent out with a new report (and a revision number) to you and the other curators for approval.

If a submission is unpublished work, it stays private. The public queue shows a claimed lineage name and the date it was claimed, and nothing else. A submission of unpublished work can be approved and hold its lineage name reservations indefinitely (i.e., until the authors publish their work and provide the additional data and publication).

The host and geography records are a separate question, and the submitter answers it. The form asks whether we may add their records now. If they say yes, the records go into the next release credited as, for example, “Ellis et al unpubl”, and are renamed to the real citation once the study is out. If they say hold, or did not answer, the records wait. Either way their lineage names are confirmed by a curator and the reservation stands, so the choice only decides when the records become public, never whether the submission is reviewed.