nale 0.1.2

Nale is a library that performs profile Hidden Markov Model (PHMM) biological sequence alignment.
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258

use crate::align::bounded::cloud_search_common::PruneStatus;
use crate::align::bounded::structs::{CloudBoundGroup, CloudMatrixLinear, CloudSearchParams, Seed};
use crate::align::bounded::{prune_and_scrub, scrub_co_located};
use crate::log_sum;
use crate::structs::{Profile, Sequence};
use crate::util::log_add;

#[inline]
pub fn compute_backward_cell(
    target: &Sequence,
    profile: &Profile,
    cloud_matrix: &mut CloudMatrixLinear,
    cloud_matrix_row_idx: usize,
    target_idx: usize,
    profile_idx: usize,
) {
    let previous_target_character = target.digital_bytes[target_idx + 1] as usize;

    //   *: the cell we are computing                     (target_idx    , profile_idx    )
    //   M: the cell of the source match state component  (target_idx + 1, profile_idx + 1)
    //   D: the cell of the source delete state component (target_idx    , profile_idx + 1)
    //   I: the cell of the source match state component  (target_idx + 1, profile_idx    )
    //
    //   classic orientation:
    //       p r o f i l e
    //     t - - - - - - - -
    //     a - - - - - - - -
    //     r - - - * D - - -
    //     g - - - I M - - -
    //     e - - - - - - - -
    //     t - - - - - - - -
    //
    //   rotated linear orientation:
    //       p r o f i l e
    //     r - - - * - - - -
    //     o - - - I D - - -
    //     w - - - - M - - -
    //
    let match_source_row_idx = (cloud_matrix_row_idx + 2) % 3;
    let insert_source_row_idx = (cloud_matrix_row_idx + 1) % 3;
    let delete_source_row_idx = (cloud_matrix_row_idx + 1) % 3;

    // match state
    //
    //  * D    >   * -
    //  I M    >   I D
    //         >   - M
    //
    cloud_matrix.set_match(
        cloud_matrix_row_idx,
        profile_idx,
        log_sum!(
            cloud_matrix.get_match(match_source_row_idx, profile_idx + 1)
                + profile.transition_score(Profile::MATCH_TO_MATCH_IDX, profile_idx)
                + profile.match_score(previous_target_character, profile_idx + 1),
            cloud_matrix.get_insert(insert_source_row_idx, profile_idx)
                + profile.transition_score(Profile::MATCH_TO_INSERT_IDX, profile_idx)
                + profile.insert_score(previous_target_character, profile_idx),
            cloud_matrix.get_delete(delete_source_row_idx, profile_idx + 1)
                + profile.transition_score(Profile::MATCH_TO_DELETE_IDX, profile_idx)
        ),
    );

    // insert state
    //
    //  * -   >   * -
    //  I M   >   I -
    //        >   - M
    //
    cloud_matrix.set_insert(
        cloud_matrix_row_idx,
        profile_idx,
        log_sum!(
            cloud_matrix.get_match(match_source_row_idx, profile_idx + 1)
                + profile.transition_score(Profile::INSERT_TO_MATCH_IDX, profile_idx)
                + profile.match_score(previous_target_character, profile_idx + 1),
            cloud_matrix.get_insert(insert_source_row_idx, profile_idx)
                + profile.transition_score(Profile::INSERT_TO_INSERT_IDX, profile_idx)
                + profile.insert_score(previous_target_character, profile_idx)
        ),
    );

    // delete state
    //
    //   * D    >    * -
    //   - M    >    - D
    //          >    - M
    //
    cloud_matrix.set_delete(
        cloud_matrix_row_idx,
        profile_idx,
        log_sum!(
            cloud_matrix.get_match(match_source_row_idx, profile_idx + 1)
                + profile.transition_score(Profile::DELETE_TO_MATCH_IDX, profile_idx)
                + profile.match_score(previous_target_character, profile_idx + 1),
            cloud_matrix.get_delete(delete_source_row_idx, profile_idx + 1)
                + profile.transition_score(Profile::DELETE_TO_DELETE_IDX, profile_idx)
        ),
    );
}

pub fn cloud_search_backward(
    profile: &Profile,
    target: &Sequence,
    seed: &Seed,
    cloud_matrix: &mut CloudMatrixLinear,
    params: &CloudSearchParams,
    bounds: &mut CloudBoundGroup,
) {
    // the highest score we've seen overall
    let mut overall_max_score = -f32::INFINITY;

    let target_end = seed.target_end.min(target.length - 1);
    let profile_end = seed.profile_end.min(profile.length - 1);

    let first_anti_diagonal_idx = target_end + profile_end;
    let gamma_anti_diagonal_idx = first_anti_diagonal_idx - params.gamma;
    let min_anti_diagonal_idx = 0usize;

    let first_cloud_matrix_row_idx = first_anti_diagonal_idx % 3;

    // setting the scores to 0 is like setting
    // the log odds ratio to 1 since log(0) = 1
    cloud_matrix.set_match(first_cloud_matrix_row_idx, profile_end, 0.0);
    cloud_matrix.set_insert(first_cloud_matrix_row_idx, profile_end, 0.0);
    cloud_matrix.set_delete(first_cloud_matrix_row_idx, profile_end, 0.0);

    // the first bound is the ending position
    bounds.set(
        first_anti_diagonal_idx,
        target_end,
        profile_end,
        target_end,
        profile_end,
    );

    for anti_diagonal_idx in (gamma_anti_diagonal_idx..first_anti_diagonal_idx).rev() {
        let previous_bound = bounds.get(anti_diagonal_idx + 1);

        bounds.set(
            anti_diagonal_idx,
            previous_bound.left_target_idx,
            previous_bound.left_profile_idx,
            previous_bound.right_target_idx,
            previous_bound.right_profile_idx,
        );

        let current_bound = bounds.get_mut(anti_diagonal_idx);
        // edge guards
        if current_bound.left_profile_idx > 1 {
            // if we haven't hit the start of the target (top row),
            // we'll plan to move the left bound up
            current_bound.left_profile_idx -= 1;
        } else {
            // otherwise we'll move to the left
            current_bound.left_target_idx -= 1;
        }

        if current_bound.right_target_idx > 1 {
            // if we haven't hit the start of the profile (first column),
            // we'll plan to move the right bound to the left
            current_bound.right_target_idx -= 1;
        } else {
            // otherwise we'll move up
            current_bound.right_profile_idx -= 1;
        }

        let cloud_matrix_row_idx = anti_diagonal_idx % 3;

        for (target_idx, profile_idx) in current_bound.anti_diagonal_cell_zip() {
            compute_backward_cell(
                target,
                profile,
                cloud_matrix,
                cloud_matrix_row_idx,
                target_idx,
                profile_idx,
            );
        }
    }

    // main recursion:
    //  - compute the next full anti-diagonal
    //  - prune with max-in-anti-diagonal rule: discard any value < max_in_current_diagonal - alpha (default: alpha = 12)
    //  - prune with x-drop rule: discard anything value < max_in_all_diagonals - beta (default: beta = 20)
    for anti_diagonal_idx in (min_anti_diagonal_idx..gamma_anti_diagonal_idx).rev() {
        let previous_bound = bounds.get(anti_diagonal_idx + 1);
        let three_forward_bound = bounds.get(anti_diagonal_idx + 3).clone();

        bounds.set(
            anti_diagonal_idx,
            previous_bound.left_target_idx,
            previous_bound.left_profile_idx,
            previous_bound.right_target_idx,
            previous_bound.right_profile_idx,
        );

        let current_bound = bounds.get_mut(anti_diagonal_idx);
        // edge guards
        if current_bound.left_profile_idx > 1 {
            // if we haven't hit the start of the target (top row),
            // we'll plan to move the left bound up
            current_bound.left_profile_idx -= 1;
        } else {
            // otherwise we'll move to the left
            current_bound.left_target_idx -= 1;
        }

        if current_bound.right_target_idx > 1 {
            // if we haven't hit the start of the profile (first column),
            // we'll plan to move the right bound to the left
            current_bound.right_target_idx -= 1;
        } else {
            // otherwise we'll move up
            current_bound.right_profile_idx -= 1;
        }

        let cloud_matrix_row_idx = anti_diagonal_idx % 3;

        // 3-forward scrub
        scrub_co_located(
            current_bound,
            &three_forward_bound,
            cloud_matrix,
            cloud_matrix_row_idx,
        );

        for (target_idx, profile_idx) in current_bound.anti_diagonal_cell_zip() {
            compute_backward_cell(
                target,
                profile,
                cloud_matrix,
                cloud_matrix_row_idx,
                target_idx,
                profile_idx,
            );
        }

        let prune_status = prune_and_scrub(
            current_bound,
            cloud_matrix,
            cloud_matrix_row_idx,
            params.alpha,
            params.beta,
            &mut overall_max_score,
        );

        match prune_status {
            PruneStatus::FullyPruned => {
                bounds.min_anti_diagonal_idx += 1;
                break;
            }
            PruneStatus::PartiallyPruned => {
                continue;
            }
        }
    }
}