Warning: This page is under construction. It will continue to change during the next few months

EMP: Enzymes and Metabolic Pathways

Status

The current volume of the EMP data includes around 2,500,000 data elements derived from 11,000 publications. Although it took over 10 years for this amount to accumulate, the encoding scheme changed very little down the way. It continues to be a mixture of formatting and tagging whose rules had never been expressed in terms of formal grammar. Until recently, visual inspection remained the only means of rule enforcement and validation.

Although the original intent was to make the data entirely machine-readable, nothing but simple keyword search techniques could match the complexity and variance in the data. The development of the parser that combines heuristics with a number of formal grammars has made it possible to recognize relations, in addition to individual facts.

Interface

The data you will be accessing from this page are organized as a number of tables stored in a SQL engine (postgreSQL). At present, there is no generic interface to these tables, other than SQL. For those not familiar with it, we will provide a number of query forms to access the most common kinds of data.

Schema

It is quite simple. Each EMP data type (not to be mixed with SQL types) is stored in a table bearing its name, one element a row. Each table has a single data attribute with the same name as the corresponding EMP type. Some numeric types also have a unit of measurement. Other attributes, common for all types, specify the location of the element in the EMP source. The only exception (so far) is the author name table that has more than one data attribute (last name, first and middle names and initials)

Use the following links to find out more about currently loaded tables (details will be added soon)

Examples:

EC Code (Custom type)

Table: ec
Name id rel loc item ec

SQL type char(12) text char(16) int2 ec_code

data SEL86357-01 NULL 29:4:29:11 1 1.6.1.1

SEL86358-17 OCN:EC:PA:SA:ECO:MM:PF:1 12:17:12:24 1 3.5.3.1

**Table: ec**
Name	id	rel	loc	item	ec
SQL type	`char(12)`	`text`	`char(16)`	`int2`	`ec_code`
data	`SEL86357-01`	`NULL`	`29:4:29:11`	`1`	`1.6.1.1`
`SEL86358-17`	`OCN:EC:PA:SA:ECO:MM:PF:1`	`12:17:12:24`	`1`	`3.5.3.1`

Attribute descriptions

id: EMP record ID
rel: Relation. If the value (here ec) was extracted from an EMP table row (relation), rel identifies that relation by providing the table header and row number
loc: Location of the element (here ec) in the original EMP record
item: Item number. If the value stored in ec originally was a ';'-separated list item, this is the element's position in the list. The default is 1 (all non-list values are actually one-element lists)
ec: The value. In this example, it is represented by the custom SQL type, ec_code. Read more about what it is useful for here. If you are a Postgres user you may be interested in its source code.

Example query

The following query can be used to retrieve all emp records containing a given EC code or a range of EC codes:

Tip: You can type virtually everything in this and other query forms on this page, like you would in an interactive SQL shell. However, selecting id, along with other attributes, will allow the query postprocessor to insert hyperlinks to the EMP records. Just for fun, you can copy the following line and paste it into the query form:

SELECT id, ini, last FROM au WHERE last = 'Jones';

KM values (Custom type)

Table: km
Name id rel loc item unit km

SQL type char(12) text char(16) int2 char(30) seg

data SEL90162-07 NULL 47:4:47:8 1 mmol/l 0.09

SEL90130-05 EF:SU:KM:2 47:23:47:43 1 umol/l 100...230

**Table: km**
Name	id	rel	loc	item	unit	km
SQL type	`char(12)`	`text`	`char(16)`	`int2`	`char(30)`	`seg`
data	`SEL90162-07`	`NULL`	`47:4:47:8`	`1`	`mmol/l`	`0.09`
`SEL90130-05`	`EF:SU:KM:2`	`47:23:47:43`	`1`	`umol/l`	`100...230`

Attribute descriptions

id, rel, loc and item are the same for all EMP types (see example above)
unit: The unit of measurement.
km: The value. In this example, it is represented by the custom SQL type, seg. If you are a Postgres user you may be interested in its source code.

Example query

The following query can be used to retrieve all emp records containing a KM that falls within the range : Just for practice, replace @ (contained) for ~ (contains) or && (overlaps) and note the difference in results. But please don't (DON'T!!!) use any other seg operator, because this is how many records will otherwise be sent yor way:

Enzyme Names (text type)

Table: en
Name id rel loc item en uc

SQL type char(12) text char(16) int2 text text

data SEL87051-05 NULL 18:4:18:48 2 glycerol-3-phosphate dehydrogenase (NAD⁺) GLYCEROL 3 PHOSPHATE DEHYDROGENASE (NAD+)

SEL87052-07 EC:EN:PA:1 24:20:24:50 2 myo-inositol 2-dehydrogenase MYO INOSITOL 2 DEHYDROGENASE

**Table: en**
Name	id	rel	loc	item	en	uc
SQL type	`char(12)`	`text`	`char(16)`	`int2`	`text`	`text`
data	`SEL87051-05`	`NULL`	`18:4:18:48`	`2`	`glycerol-3-phosphate dehydrogenase (NAD⁺)`	`GLYCEROL 3 PHOSPHATE DEHYDROGENASE (NAD+)`
`SEL87052-07`	`EC:EN:PA:1`	`24:20:24:50`	`2`	`myo-inositol 2-dehydrogenase`	`MYO INOSITOL 2 DEHYDROGENASE`

Attribute descriptions

id, rel, loc and item are the same for all EMP types (see examples above)
en: Enzyme name -- as recommended by the Enzyme Nomenclature.
uc: Upper-cased and simplified en. In the real life, enzyme names may contain various font and formatting sequences; these are eliminated from uc in order to simplify indexing.

Example query

The following query can be used to retrieve all emp records containing a KM that falls within the range :

Gene Selkov Jr.

Last modified: Fri May 1 22:33:50 MSD 1998