MEME5 Meeting Notes

Owned by jo (Unlicensed)
Last updated: Sept 08, 2017 by Rick Wood

Table of Contents:

Agenda for Upcoming Meeting

September 8, 2017, 1:30PM

  • NCI insertion: finished, taking 14 hours to complete.   
    • Laura findings:

      • I see a couple of these… they have a demoted atom and in demotions bin but there is no 2nd concept or demotion rel.

        • CN# 101298 One CUI C0205447
        • TODO: look into

      • The PAR and CHD are wrong.

        What is listed as CHD are actually the PAR of  this concept. If you look at context for SNOMED for example, Finding of HIV status is a PAR.

        • CN# 262978 HIV negative finding CUI C0481430

        • TODO: deploy code updates

      • For the list of checklists, can we have it that the newest ones created are on the 1st page? So the order is newest to oldest?

        • Not a sorting issue, but an insertion when creating.  Also,  feel this is lower priority.

        • TODO: refresh page after creation of new checklist

  • Post-query processing for concepts_id_1,concepts_id_2 bins.
    • An issue for QA bins, and for demotions ME bin
    • Example: after MTH_2017AA insertion, it took 42620.183 seconds (12 hours) to process ME 'demotions'. So no cycles.
      • 16,152 conceptId pairs returned by query, 1994 clusters ended up in bin
        • 6653237 - 981810
        • 981810 - 6653237
        • 682110 - 6653664
        • 6653664 - 682110
      • 3716 entries in parChd.
        • Parent 6660138 went through 130+ iterations of getDescendents
        • 6660138=[6660140,6660141,6660142,6660143,6660144,6660145,6660146]
        • 6660140=[6660141,6660142,6660143,6660144,6660145,6660146]
        • 6660141=[6660142,6660143,6660144,6660145,6660146]
        • 6660142=[6660143,6660144,6660145,6660146]
        • 6660143=[6660144,6660145,6660146]
        • 6660144=[6660145,6660146]
        • 6660145=[6660146]
        • 6660146=[]
      • Checking to make sure child hasn't already been processed before running
  • Process export/import issue
    • Export updated so JSON contains parameters
    • Import needs to be able to incorporate JSON parameters into algorithm properties
  • MTH_2017AA insertion
    • Still no NEEDS_REVIEW checklists
      • Precomputed Merges - all lines set to N|N
      • Generated Merges - changeStatus set to false
    • MEME4 vs. MEME5 merge scripts review

September 1, 2017, 1:30PM

  • QA Bins: post-query processing for concepts_id_1,concepts_id_2 bins never finishes
  • Continue term type suppressability discussion

    • Per the UMLS reference manual:

      Suppressible flag. Values = O, E, Y, or N

      O: All obsolete content, whether they are obsolesced by the source or by NLM. These will include all atoms having obsolete TTYs, and other atoms becoming obsolete that have not acquired an obsolete TTY (e.g. RxNorm SCDs no longer associated with current drugs, LNC atoms derived from obsolete LNC concepts).

      E: Non-obsolete content marked suppressible by an editor. These do not have a suppressible SAB/TTY combination.

      Y: Non-obsolete content deemed suppressible during inversion. These can be determined by a specific SAB/TTY combination explicitly listed in MRRANK.

      N: None of the above

    • Per Stephanie:

      I am  looking at cases where tty = AB or HS && suppressible != “Y”.     For most of them, a simple change to NCI & ISO3166-2 would do it.  I think if we explain it to NCI they will ok it -- Laura will understand, and we should probably run it by Tracey/Rob/Larry in case they have any concerns or foresee any gotchas.  For the others that come in through UMLS … I think the but a few come from RxNorm which has its own rules so I’m pretty sure they’re intentional and I’m hesitant to change it.   I would be ok changing the two “HS” cases to “Y” but then it would be different from the UMLS.  Not sure if that matters.

  • Review umlscui.txt issues

Action Items for Future Meetings

  • tba

Notes from Previous Meetings

August 25, 2017

  • Problems with XR relationships
  • Reversed RELAs
  • Suppressibility
  • Legacy Codes
  • Expanded Hierarchy
  • Ambiguous Flags in Concept Reports