annotate libs/sratoolkit.2.8.0-centos_linux64/schema/align/refseq.vschema @ 3:38ad1130d077 draft

planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
author charles_s_test
date Mon, 27 Nov 2017 11:21:07 -0500
parents
children
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
3
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
1 /*===========================================================================
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
2 *
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
3 * PUBLIC DOMAIN NOTICE
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
4 * National Center for Biotechnology Information
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
5 *
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
6 * This software/database is a "United States Government Work" under the
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
7 * terms of the United States Copyright Act. It was written as part of
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
8 * the author's official duties as a United States Government employee and
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
9 * thus cannot be copyrighted. This software/database is freely available
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
10 * to the public for use. The National Library of Medicine and the U.S.
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
11 * Government have not placed any restriction on its use or reproduction.
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
12 *
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
13 * Although all reasonable efforts have been taken to ensure the accuracy
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
14 * and reliability of the software and data, the NLM and the U.S.
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
15 * Government do not and cannot warrant the performance or results that
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
16 * may be obtained by using this software or data. The NLM and the U.S.
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
17 * Government disclaim all warranties, express or implied, including
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
18 * warranties of performance, merchantability or fitness for any particular
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
19 * purpose.
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
20 *
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
21 * Please cite the author in any work or product based on this material.
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
22 *
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
23 * ===========================================================================
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
24 *
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
25 */
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
26
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
27 /*==========================================================================
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
28 * VDB Reference Sequence types, functions and tables
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
29 */
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
30 version 1;
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
31
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
32 include 'vdb/vdb.vschema';
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
33 include 'ncbi/seq.vschema';
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
34 include 'ncbi/sra.vschema';
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
35 include 'ncbi/stats.vschema';
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
36
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
37
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
38 extern function U8 NCBI:refseq:stats #2 ( INSDC:4na:bin seq, INSDC:coord:len len )
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
39 = NCBI:refSeq:stats;
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
40
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
41 table NCBI:refseq:tbl:reference #1.0.2 =
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
42 NCBI:tbl:base_space #2.0.2,
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
43 NCBI:tbl:phred_quality #2.0.3,
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
44 NCBI:tbl:seqloc #1.0,
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
45 NCBI:SRA:tbl:stats #1.1.2
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
46 {
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
47 // 128K
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
48 column default limit = 131072;
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
49
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
50 extern column U32 MAX_SEQ_LEN; /* must be static */
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
51 extern column < ascii > izip_encoding DEF_LINE; /* remainder of defline after SEQ_ID */
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
52
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
53 // trigger upconverts to INSDC:dna:text to get MD5
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
54 trigger table_stats
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
55 = NCBI:refseq:stats(in_4na_bin, _alt_in_read_len);
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
56
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
57 readonly column U64 TOTAL_SEQ_LEN
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
58 = < U64 > meta:value < 'STATS/TOTAL_SEQ_LEN', true >();
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
59
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
60 readonly column U8[16] MD5
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
61 = < U8[16] > meta:read < 'STATS/MD5', true >();
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
62
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
63 // indicates if sequence has circular structure
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
64 // should be static
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
65 extern column bool_encoding CIRCULAR;
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
66
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
67 /* columns:
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
68 * READ
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
69 * QUALITY (optional)
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
70 * SEQ_ID
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
71 * SEQ_START
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
72 * SEQ_LEN
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
73 * MAX_SEQ_LEN
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
74 * TOTAL_SEQ_LEN
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
75 * DEF_LINE
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
76 */
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
77
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
78 // make CS_KEY writable
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
79 INSDC:dna:text in_cs_key
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
80 = < INSDC:dna:text, INSDC:dna:text > map < 'acgtn', 'ACGTN' > ( CS_KEY );
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
81 physical column < INSDC:dna:text > zip_encoding .CS_KEY = in_cs_key;
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
82 // extra columns needed for CS conversion
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
83 INSDC:coord:zero out_read_start = < INSDC:coord:zero> echo < 0 > ();
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
84 INSDC:coord:len out_read_len = .SEQ_LEN;
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
85
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
86 INSDC:coord:len _alt_in_read_len
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
87 = READ_LEN
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
88 | SEQ_LEN;
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
89
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
90 INSDC:SRA:xread_type _alt_in_read_type
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
91 = READ_TYPE
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
92 | < INSDC:SRA:xread_type > echo < SRA_READ_TYPE_BIOLOGICAL > ();
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
93
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
94 INSDC:SRA:xread_type out_read_type
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
95 = .READ_TYPE
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
96 | < INSDC:SRA:xread_type > echo < SRA_READ_TYPE_BIOLOGICAL > ();
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
97 };
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
98
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
99 // older spelling
38ad1130d077 planemo upload commit a4fb57231f274270afbfebd47f67df05babffa4a-dirty
charles_s_test
parents:
diff changeset
100 alias NCBI:refseq:tbl:reference NCBI:refSeq:tbl:reference;