Skip to content

Commit 402de74

Browse files
pd3pd3
authored
Merge pull request #1942 from tecnickcom/develop
Update VariantKey with latest changes.
2 parents 4d0d262 + 99e9756 commit 402de74

5 files changed

Lines changed: 66 additions & 36 deletions

File tree

‎LICENSE‎

Lines changed: 2 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -723,11 +723,12 @@ Public License instead of this License. But first, please read
723723

724724
-----------------------------------------------------------------------------
725725

726-
LICENSE FOR VariantKey (https://github.com/Genomicsplc/variantkey)
726+
LICENSE FOR VariantKey (https://github.com/tecnickcom/variantkey)
727727

728728
The MIT License
729729

730730
Copyright (c) 2017-2018 GENOMICS plc
731+
Copyright (c) 2018-2023 Nicola Asuni - Tecnick.com
731732

732733
Permission is hereby granted, free of charge, to any person obtaining a copy
733734
of this software and associated documentation files (the "Software"), to deal

‎hex.h‎

Lines changed: 3 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -3,10 +3,10 @@
33
// hex.h
44
//
55
// @category Libraries
6-
// @author Nicola Asuni <nicola.asuni@genomicsplc.com>
6+
// @author Nicola Asuni <info@tecnick.com>
7+
// @link https://github.com/tecnickcom/variantkey
8+
// @license MIT [LICENSE](https://raw.githubusercontent.com/tecnickcom/variantkey/main/LICENSE)
79
// @copyright 2017-2018 GENOMICS plc
8-
// @license MIT (see LICENSE)
9-
// @link https://github.com/genomicsplc/variantkey
1010
//
1111
// LICENSE
1212
//

‎plugins/add-variantkey.c‎

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -2,7 +2,7 @@
22
33
Copyright (C) 2017-2018 GENOMICS plc.
44
5-
Author: Nicola Asuni <nicola.asuni@genomicsplc.com>
5+
Author: Nicola Asuni <nicola.asuni@tecnick.com>
66
77
Permission is hereby granted, free of charge, to any person obtaining a copy
88
of this software and associated documentation files (the "Software"), to deal

‎plugins/variantkey-hex.c‎

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -2,7 +2,7 @@
22
33
Copyright (C) 2017-2018 GENOMICS plc.
44
5-
Author: Nicola Asuni <nicola.asuni@genomicsplc.com>
5+
Author: Nicola Asuni <nicola.asuni@tecnick.com>
66
77
Permission is hereby granted, free of charge, to any person obtaining a copy
88
of this software and associated documentation files (the "Software"), to deal

‎variantkey.h‎

Lines changed: 59 additions & 30 deletions
Original file line numberDiff line numberDiff line change
@@ -3,14 +3,15 @@
33
// variantkey.h
44
//
55
// @category Libraries
6-
// @author Nicola Asuni <nicola.asuni@genomicsplc.com>
7-
// @copyright 2017-2018 GENOMICS plc
8-
// @license MIT (see LICENSE)
9-
// @link https://github.com/genomicsplc/variantkey
6+
// @author Nicola Asuni <info@tecnick.com>
7+
// @link https://github.com/tecnickcom/variantkey
8+
// @license MIT [LICENSE](https://raw.githubusercontent.com/tecnickcom/variantkey/main/LICENSE)
9+
// @copyright 2017-2018 GENOMICS plc, 2018-2023 Nicola Asuni - Tecnick.com
1010
//
1111
// LICENSE
1212
//
1313
// Copyright (c) 2017-2018 GENOMICS plc
14+
// Copyright (c) 2018-2023 Nicola Asuni - Tecnick.com
1415
//
1516
// Permission is hereby granted, free of charge, to any person obtaining a copy
1617
// of this software and associated documentation files (the "Software"), to deal
@@ -54,6 +55,7 @@
5455
#define VKMASK_REFALT 0x000000007FFFFFFF //!< VariantKey binary mask for REF+ALT [ 00000000 00000000 00000000 00000000 01111111 11111111 11111111 11111111 ]
5556
#define VKSHIFT_CHROM 59 //!< CHROM LSB position from the VariantKey LSB
5657
#define VKSHIFT_POS 31 //!< POS LSB position from the VariantKey LSB
58+
#define MAXUINT32 0xFFFFFFFF //!< Maximum value for uint32_t
5759

5860
/**
5961
* VariantKey struct.
@@ -75,16 +77,54 @@ typedef struct vkrange_t
7577
uint64_t max; //!< Maximum VariantKey value for any given REF+ALT encoding
7678
} vkrange_t;
7779

78-
/** @brief Returns chromosome numerical encoding.
80+
/** @brief Returns the encoding for a numerical chromosome input.
7981
*
8082
* @param chrom Chromosome. An identifier from the reference genome, no white-space permitted.
8183
* @param size Length of the chrom string, excluding the terminating null byte.
8284
*
8385
* @return CHROM code
8486
*/
87+
static inline uint8_t encode_numeric_chrom(const char *chrom, size_t size)
88+
{
89+
size_t i;
90+
uint8_t v = (chrom[0] - '0');
91+
for (i = 1; i < size; i++)
92+
{
93+
if ((chrom[i] > '9') || (chrom[i] < '0'))
94+
{
95+
return 0; // NA: a character that is not a numebr was found.
96+
}
97+
v = ((v * 10) + (chrom[i] - '0'));
98+
}
99+
return v;
100+
}
101+
102+
103+
/** @brief Returns a true value (1) if the input chrom has 'chr' prefix (case insensitive).
104+
*
105+
* @param chrom Chromosome. An identifier from the reference genome, no white-space permitted.
106+
* @param size Length of the chrom string, excluding the terminating null byte.
107+
*
108+
* @return True (1) if the chr prefix is present.
109+
*/
110+
static inline int has_chrom_chr_prefix(const char *chrom, size_t size)
111+
{
112+
return ((size > 3)
113+
&& ((chrom[0] == 'c') || (chrom[0] == 'C'))
114+
&& ((chrom[1] == 'h') || (chrom[1] == 'H'))
115+
&& ((chrom[2] == 'r') || (chrom[2] == 'R')));
116+
}
117+
118+
/** @brief Returns chromosome numerical encoding.
119+
*
120+
* @param chrom Chromosome. An identifier from the reference genome, no white-space permitted.
121+
* @param size Length of the chrom string, excluding the terminating null byte.
122+
*
123+
* @return CHROM code or 0 in case of invalid input.
124+
*/
85125
static inline uint8_t encode_chrom(const char *chrom, size_t size)
86126
{
87-
// X > 23 ; Y > 24 ; M > 25
127+
// X = 23; Y = 24; M = 25; any other letter is mapped to 0:
88128
static const uint8_t onecharmap[] =
89129
{
90130
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
@@ -98,32 +138,19 @@ static inline uint8_t encode_chrom(const char *chrom, size_t size)
98138
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
99139
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
100140
};
101-
// remove "chr" prefix
102-
if ((size > 3)
103-
&& ((chrom[0] == 'c') || (chrom[0] == 'C'))
104-
&& ((chrom[1] == 'h') || (chrom[1] == 'H'))
105-
&& ((chrom[2] == 'r') || (chrom[2] == 'R')))
141+
if (has_chrom_chr_prefix(chrom, size))
106142
{
143+
// remove "chr" prefix
107144
chrom += 3;
108145
size -= 3;
109146
}
110147
if (size == 0)
111148
{
112149
return 0;
113150
}
114-
if ((chrom[0] <= '9') && (chrom[0] >= '0')) // Number
151+
if ((chrom[0] <= '9') && (chrom[0] >= '0'))
115152
{
116-
size_t i;
117-
uint8_t v = (chrom[0] - '0');
118-
for (i = 1; i < size; i++)
119-
{
120-
if ((chrom[i] > '9') || (chrom[i] < '0'))
121-
{
122-
return 0; // NA
123-
}
124-
v = ((v * 10) + (chrom[i] - '0'));
125-
}
126-
return v;
153+
return encode_numeric_chrom(chrom, size);
127154
}
128155
if ((size == 1) || ((size == 2) && ((chrom[1] == 'T') || (chrom[1] == 't'))))
129156
{
@@ -159,10 +186,10 @@ static inline uint32_t encode_base(const uint8_t c)
159186
{
160187
/*
161188
Encode base:
162-
A > 0
163-
C > 1
164-
G > 2
165-
T > 3
189+
A = 0
190+
C = 1
191+
G = 2
192+
T = 3
166193
*/
167194
static const uint32_t map[] =
168195
{
@@ -205,7 +232,7 @@ static inline uint32_t encode_refalt_rev(const char *ref, size_t sizeref, const
205232
uint8_t bitpos = 23;
206233
if ((encode_allele(&h, &bitpos, ref, sizeref) < 0) || (encode_allele(&h, &bitpos, alt, sizealt) < 0))
207234
{
208-
return 0; // error code
235+
return MAXUINT32; // error code
209236
}
210237
return h;
211238
}
@@ -318,7 +345,7 @@ static inline uint32_t encode_refalt(const char *ref, size_t sizeref, const char
318345
if ((sizeref + sizealt) <= 11)
319346
{
320347
uint32_t h = encode_refalt_rev(ref, sizeref, alt, sizealt);
321-
if (h != 0)
348+
if (h != MAXUINT32)
322349
{
323350
return h;
324351
}
@@ -486,7 +513,9 @@ static inline void decode_variantkey(uint64_t code, variantkey_t *vk)
486513
vk->refalt = extract_variantkey_refalt(code);
487514
}
488515

489-
/** @brief Returns a 64 bit variant key based on CHROM, POS (0-based), REF, ALT.
516+
/**
517+
* Returns a 64 bit variant key based on CHROM, POS (0-based), REF, ALT.
518+
* The variant should be already normalized (see normalize_variant or use normalized_variantkey).
490519
*
491520
* @param chrom Chromosome. An identifier from the reference genome, no white-space or leading zeros permitted.
492521
* @param sizechrom Length of the chrom string, excluding the terminating null byte.

0 commit comments

Comments
 (0)