Mailing-List: contact dev-help@ctakes.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@ctakes.apache.org
Received-SPF: softfail (nike.apache.org: transitioning domain of psz@mit.edu
 does not designate 128.30.2.149 as permitted sender)
Content-Type: text/plain; charset=utf-8
Mime-Version: 1.0 (Mac OS X Mail 7.3 \(1878.6\))
Subject: Re: to map UMLS CUI with normalized form
From: Peter Szolovits <psz@mit.edu>
In-Reply-To: 
 <CA+VTBdFE00S-ts4y0YqLT+HC-nSVfJ3Xi-6uw7sC7auUDoWUkA@mail.gmail.com>
Date: Thu, 4 Sep 2014 07:34:07 -0400
Content-Transfer-Encoding: quoted-printable
Message-Id: <42D64FF4-1DC2-42E4-805D-A5FD968EA048@mit.edu>
References: 
 <CA+VTBdHUwegUpP-uas0PnQxd_SvfKv5GaQkG=8ApxyyveqONOA@mail.gmail.com>
 <D65CD179-63CE-4128-88B9-7218009BBB8E@childrens.harvard.edu>
 <CA+VTBdFNoMvhz04-YPARfGhD3TVDBNvjgck4ipHJT5b2fMueDQ@mail.gmail.com>
 <B107279F-575E-4AC9-BD7A-DD4697E7A2BB@mit.edu>
 <CA+VTBdFE00S-ts4y0YqLT+HC-nSVfJ3Xi-6uw7sC7auUDoWUkA@mail.gmail.com>
To: dev@ctakes.apache.org

You need to (a) get a (free) license to use UMLS, then (b) download the =
(large) distribution, and (c) install it in a local database.  See =
https://uts.nlm.nih.gov

On Sep 4, 2014, at 5:33 AM, Prakash Poudyal <prakashpoudyal@gmail.com> =
wrote:

> Hi Peter,
>=20
> Thanks for wonderful. Can you tell me how to get this database.
>=20
>=20
> On Mon, Sep 1, 2014 at 10:32 PM, Peter Szolovits <psz@mit.edu> wrote:
>=20
>> A single CUI may have many different preferred names in different
>> vocabularies.  If you have a mysql version of UMLS installed, you can =
do
>> something like
>> CREATE VIEW pname AS
>> select cui, lat, ts, lui, stt, sui, ispref, aui, saui, scui, sdui, =
sab,
>> tty, code, str, srl, suppress, cvf
>> from `mrconso`
>> where (ts =3D 'P')
>> and (stt =3D 'PF')
>> and (ispref =3D 'Y')
>> and (lat =3D 'ENG'))
>>=20
>> to define the preferred terms for each CUI.  I.e., this gives a =
subset of
>> MRCONSO containing only the preferred name of each concept.  As you =
can see
>> from the table below, some vocabularies provide many more of these =
than
>> others.  Then, to find that preferred term for a concept, you can =
just do
>> something like:
>> mysql> select str from pname where cui=3D'C1141949';
>> +----------------------+
>> | STR                  |
>> +----------------------+
>> | Troponin I increased |
>> +----------------------+
>> 1 row in set (0.00 sec)
>>=20
>> but
>>=20
>> mysql> select distinct str from mrconso where cui=3D'C1141949';
>> +---------------------------------+
>> | str                             |
>> +---------------------------------+
>> | Troponin I zv=C3=BD=C5=A1en=C3=BD              |
>> | troponine I verhoogd            |
>> | Troponin I increased            |
>> | Troponine I augment=C3=A9e           |
>> | Troponin I erhoeht              |
>> | Troponin I emelkedett           |
>> | Troponina I aumentata           |
>> | =E3=83=88=E3=83=AD=E3=83=9D=E3=83=8B=E3=83=B3=EF=BC=A9=E5=A2=97=E5=8A=
=A0                |
>> | =EF=BE=84=EF=BE=9B=EF=BE=8E=EF=BE=9F=EF=BE=86=EF=BE=9DI=EF=BD=BF=EF=BE=
=9E=EF=BD=B3=EF=BD=B6                     |
>> | Troponina I aumentada           |
>> +---------------------------------+
>> 10 rows in set (0.00 sec)
>>=20
>> In this case, the source of ambiguity is only from different =
languages,
>> but it could also be from the same CUI appearing in different SABs
>> (vocabularies) with different names.
>>=20
>> mysql> select sab, count(*) c from pname group by sab order by c desc
>> limit 40;
>> +---------------+--------+
>> | SAB           | c      |
>> +---------------+--------+
>> | NCBI          | 788995 |
>> | MSH           | 308591 |
>> | MEDCIN        | 253291 |
>> | SNOMEDCT      | 228656 |
>> | RXNORM        | 220148 |
>> | ICD10PCS      | 178093 |
>> | MTH           | 157083 |
>> | LNC           | 131930 |
>> | ICD10CM       |  81082 |
>> | FMA           |  72645 |
>> | GO            |  57326 |
>> | OMIM          |  46058 |
>> | NCI           |  40962 |
>> | RCD           |  34638 |
>> | MDR           |  24144 |
>> | ICPC2ICD10ENG |  23618 |
>> | MMX           |  22387 |
>> | CPT           |  20064 |
>> | MMSL          |  19419 |
>> | UMD           |  15386 |
>> | NDDF          |  14635 |
>> | VANDF         |  12483 |
>> | SNMI          |  11222 |
>> | NIC           |  10487 |
>> | MTHSPL        |   9978 |
>> | NDFRT         |   9764 |
>> | ICD10AM       |   9098 |
>> | CCPSS         |   8226 |
>> | MTHFDA        |   6545 |
>> | ICD9CM        |   6521 |
>> | AOD           |   6513 |
>> | RCDSY         |   6275 |
>> | HCPCS         |   5200 |
>> | HL7V3.0       |   5097 |
>> | PDQ           |   4941 |
>> | MDDB          |   4938 |
>> | MTHICD9       |   4721 |
>> | CSP           |   3793 |
>> | GS            |   3770 |
>> | NOC           |   3645 |
>> +---------------+--------+
>> 40 rows in set (5.31 sec)
>>=20
>> On Sep 1, 2014, at 4:37 PM, Prakash Poudyal =
<prakashpoudyal@gmail.com>
>> wrote:
>>=20
>>> Hi Chen,
>>>=20
>>> Thanks for mail.  I may be wrong,
>>>=20
>>> dizziness (normalized form) =3D C0002940 (CUI value)
>>>=20
>>> I am searching a system in which if I enter C002940 than dizziness =
could
>>> come. Or is there any index, or dictionary for it.
>>>=20
>>> If you don't understand please write me again.
>>>=20
>>> Thanks
>>>=20
>>> Regards
>>> Prakash
>>>=20
>>>=20
>>> On Mon, Sep 1, 2014 at 9:13 PM, Chen, Pei <
>> Pei.Chen@childrens.harvard.edu>
>>> wrote:
>>>=20
>>>> Hi Prakash,
>>>> Could you clarify what you mean by 'normalized form'?  An example?
>>>>=20
>>>> -Pei
>>>>=20
>>>> Sent from my iPhone
>>>>=20
>>>>> On Sep 1, 2014, at 9:16 AM, "Prakash Poudyal" <
>> prakashpoudyal@gmail.com>
>>>> wrote:
>>>>>=20
>>>>> Hi!
>>>>>=20
>>>>> I am working in cTAKES to analysis clinical document. Is it =
possible
>> for
>>>> me
>>>>> to know how the CUI code is provided to the normalized form.
>>>>>=20
>>>>> Is there any dictionary or webportal that could map  SONMED CT =
UMLS CUI
>>>>> code with normalized form.
>>>>>=20
>>>>> --
>>>>>=20
>>>>> Regards
>>>>> Prakash Poudyal
>>>>=20
>>>=20
>>>=20
>>>=20
>>> --
>>>=20
>>> Regards
>>> Prakash Poudyal
>>=20
>>=20
>=20
>=20
> --=20
>=20
> Regards
> Prakash Poudyal