Gforge Advanced Server - Screensaver HTS LIMS - Screensaver Tasks

viewing of mutual positives data from others' screen viewer page

if a user can view mutual positives of another's screen result, then the user should be allowed to view the screen result from other's screen viewer page, where screen result data is limited to the subset of mutual positives (currently, the only way to view another's mutual positives is to add them in to a well search results as a new column).

assay wells 'shareable' flag

We have use cases where subsets of screen results (sets of wells) need to be flagged as public or private, in order to satisfy screeners whose publications only reveal subsets of screen result data. We could implement by adding a flag to AssayWell and and populate this based upon a new column in our screen result file format (a fixed column, not a data header).

re-profile screen result parser memory optimizations

On the data-sharing-levels branch, the SRP has been modified to make use of the new AssayWell entity. We need to profile the SRP again to ensure our memory optimizations made in 1.8.1-dev are still in effect.

Pubchem PUG server fails to complete for queries

It appears that _all_ queries to the PUG server are failing to return results. Here is the text of my email to NIH: Hi, I am trying to use the http://pubchem.ncbi.nlm.nih.gov/pug/pug.cgi server to access molecule information from your database, however, I am finding that all queries are failing to complete, even queries that I have successfully run using the http://pubchem.ncbi.nlm.nih.gov/search/search.cgi interface, In the following test case, I am using one such successful query, and I am using the "save query" option to get the following XML which I am then attempting to submit to the pug interface: Saved Query:

ClC3=C(\C=C/C1=C(N(N(C1=O)C2=CC=CC=C2)C)C)C=CC(=C3)Cl

2000000

When I submit this to the pug server, I get the following response, with a "reqid": $curl -d @pc_structure_search2.xml "http://pubchem.ncbi.nlm.nih.gov/pug/pug.cgi"

545074283025045613

Structure search job was submitted

for which I construct the following status request in order to get my results, if and when completed:

545074283025045613

but I am always getting the response: $ curl -d @pc_status2.xml "http://pubchem.ncbi.nlm.nih.gov/pug/pug.cgi"

Your search is running, please wait...

545074283025045613

Your search is running, please wait...

I am wondering if the pug.cgi server is currently working and if so, what could be the problem with my queries? Does anyone know a better contact at PubChem with whom we could consult to get at the issue here?

delete temp files used for batch I/O immediately after use

These files can be large, and can consume considerable amounts of space. They will only be cleaned up automatically on web app restarts, which is too infrequent.

marcus admins should only be able to edit marcus checklist items

Caroline has requested that Marcus admins only be able to edit marcus checklist items, if they have the userChecklistItemsAdmin role

query for PMCID using PMID and vice versa

In Screen Viewer, when adding a new publication, the user should be able to enter the PMID or PMCID and obtain all of the publiation information from Entrez for either ID, including the other ID, if possible.

report duplicate reagent well error on library contents import

Currently, these errors manifest as system errors at database flush time, and are only indicated in log files.

create AssayWell entity

This will allow us to normalize attributes we currently store in ResultValue entities, including assayWellType and isPositive. We will store isPositive at the AssayWell level, indicating whether *any* result value for the assay well has been flagged as positive. In particular, it has been empirically determined that explicit storage of this aggregate value will allow Data Sharing Level 2 (mutual positives screen sharing) to be computed efficiently. AssayWell should also play a critical part in improving the performance of WellSearchResults

rename "screensaverUser" role

Rename to "login access" (or something similar), to make role's intent more clear.

concurrent editing error incorrect

When an entity is simultaneously edited by multiple users (causing a concurrent modification exception), the error is being handled as a system error, rather than as a less-critical application error. That is, the error message is displayed at the top page (correctly), but is being shown on the "exception" reporting page, rather than reloading the entity viewer page with the latest version of the entity. This was observed with the Screen Viewer, but probably affects all entity viewers.

inform user of login failure reason

This will reduce the administrative overhead of dealing with users' "I cannot login" support requests. In particular, it should be pointed out that user agreements need to be signed if they have not been.

CPR quick finder fails when using legacy visit number

Generates Hibernate error

implement HTML formatting of molecular formulas

MolecularFormula.toHtml() should generate HTML that formats the formula using subscript and superscript output, as appropriate.

auto-(un)subscribe users from listserv mailing lists based upon checklist items

Many user checklist items track whether a user is a member of a listserv mailing list. Rather than using this to simply track their membership, and requiring after-the-fact data entry into Screensaver, we should use Screensaver's checklist items to automatically update the listserv membership lists. It is acceptable to implement this as a batch process that mirrors the users' status over to the listserv membership. This does not have to be part of Screensaver proper.

add user checklist items for Pipeline Pilot access lists

We will need 2 need checklist items to track users membership on Pipeline Pilot access lists.

auto-add screener facility usage roles when user associated with a screen

Small molecule user and RNAi user roles should be added automatically when a screen of a particular type is associated with a user

change "time point" RVT field to numeric type

The RVT "time point" property is a string, which makes its data somewhat useless for analysis purposes. We should convert this to a unitized numeric field. At ICCB-L all data is in hour units, but we may want a finer resolution (minutes, seconds?) for increased flexibility. We should also enforce that this time point is a relative, elapsed time (from the time of transfection, perhaps?). Cor Lieftink (NKI) is interested in possibly using this field for his facility's high-content data, but expressed some concerns, in particular the lack of numeric typing.

screen result file format should allow data header attributes to be optional

Not all data header attributes ("channel", "time point", et. al.) are needed by all facilities. To reduce effort, the screen result parser should find data header attributes by name, rather than fixed row index, allowing unused attributes to be left out of a screen result import file. This also allows new attributes to be added without "breaking" existing screen result files (they can still be imported despite the missing attributes).

PubChem submissions

Screensaver should aid the facility in submitting (published) screen result data to PubChem. Not sure what this involves currently, but there must be export formats and/or conversions that can be automatically performed by Screensaver, if not the full submission itself.