poasta 0.1.0

Fast, optimal, gap-affine partial order 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

use std::cmp::{max, min};
use std::collections::BTreeSet;
use std::marker::PhantomData;
use std::ops::{RangeFrom, RangeTo};
use crate::aligner::offsets::OffsetType;
use crate::aligner::scoring::{AlignmentCosts, GetAlignmentCosts, Score};
use crate::aligner::aln_graph::{AlignmentGraphNode, AlignState};
use crate::aligner::astar::AstarVisited;
use crate::bubbles::index::{BubbleIndex, NodeBubbleMap};
use crate::graphs::NodeIndexType;


pub struct ReachedBubbleExitsMatch<'a, V, N, O>
    where V: AstarVisited<N, O> + GetAlignmentCosts,
          N: NodeIndexType,
          O: OffsetType,
{
    visited: &'a V,
    reached_offsets: &'a BTreeSet<O>,
    seq_len: usize,
    dummy: PhantomData<N>,
}

impl<'a, V, N, O> ReachedBubbleExitsMatch<'a, V, N, O>
    where V: AstarVisited<N, O> + GetAlignmentCosts,
          N: NodeIndexType,
          O: OffsetType,
{
    pub fn new(visited: &'a V, reached_offsets: &'a BTreeSet<O>, seq_len: usize) -> Self {
        Self {
            visited,
            reached_offsets,
            seq_len,
            dummy: PhantomData,
        }
    }

    pub fn can_improve_bubble(
        &self, bubble_index: &BubbleIndex<N>,
        bubble: &NodeBubbleMap<N>,
        aln_node: &AlignmentGraphNode<N, O>,
        aln_state: AlignState,
        current_score: &Score
    ) -> bool {
        if self.reached_offsets.is_empty() {
            return true;
        }

        if aln_node.node() == bubble.bubble_exit {
            return true;
        }

        // j^min and j^max
        let target_offset_min = aln_node.offset() + O::new(bubble.min_dist_to_exit);
        let target_offset_max = aln_node.offset() + O::new(bubble.max_dist_to_exit);

        let min_dist_to_end = bubble_index.get_min_dist_to_end(bubble.bubble_exit)
            .saturating_sub(1);

        if target_offset_max.as_usize() > self.seq_len {
            return true;
        }

        // eprintln!("Checking for improving over bubble, [o, o'^min, o'^max] = [{:?}, {target_offset_min:?}, {target_offset_max:?}]", aln_node.offset());
        // eprintln!("Bubble exit: {:?}", bubble.bubble_exit);

        let mut prev_reached = self.reached_offsets
            .range(RangeTo { end: target_offset_min }).next_back();

        let mut last_offset = None;

        for next_reached in self.reached_offsets
            .range(target_offset_min..=target_offset_max)
        {
            let offset1 = prev_reached.map_or(
                target_offset_min,
                |v| max(target_offset_min, *v + O::one())
            );

            // Gap-affine and 2-piece alignment costs only.
            // If already in a deletion or insertion state, we need to take into account that
            // the current state under test would not have to incur the gap open cost again,
            // while an implicitly opened gap from a reached Match state would.
            if aln_state == AlignState::Deletion {
                let next_reached_cost = self.visited.get_score(&AlignmentGraphNode::new(
                    bubble.bubble_exit,
                    *next_reached
                ), AlignState::Match);

                if next_reached_cost + self.visited.get_costs().gap_open() > *current_score {
                    return true
                }
            } else if aln_state == AlignState::Deletion2 {
                let next_reached_cost = self.visited.get_score(&AlignmentGraphNode::new(
                    bubble.bubble_exit,
                    *next_reached
                ), AlignState::Match);

                if next_reached_cost + self.visited.get_costs().gap_open2() > *current_score {
                    return true
                }
            }

            if let Some(prev) = prev_reached {
                if aln_state == AlignState::Insertion {
                    let prev_reached_cost = self.visited.get_score(&AlignmentGraphNode::new(
                        bubble.bubble_exit,
                        *prev
                    ), AlignState::Match);

                    if prev_reached_cost + self.visited.get_costs().gap_open() > *current_score {
                        return true
                    }
                } else if aln_state == AlignState::Insertion2 {
                    let prev_reached_cost = self.visited.get_score(&AlignmentGraphNode::new(
                        bubble.bubble_exit,
                        *prev
                    ), AlignState::Match);

                    if prev_reached_cost + self.visited.get_costs().gap_open2() > *current_score {
                        return true
                    }
                }
            }

            // eprintln!("Checking o1... {offset1:?} (at score {current_score:?}");
            if self
                .can_improve_at_offset(bubble.bubble_exit, offset1, current_score, prev_reached, Some(next_reached), min_dist_to_end)
            {
                return true;
            }

            let offset2 = min(target_offset_max, max(target_offset_min, *next_reached - O::one()));
            if offset2 != offset1 {
                // eprintln!("Checking o2... {offset2:?} (at score {current_score:?}");

                if self
                    .can_improve_at_offset(bubble.bubble_exit, offset2, current_score, prev_reached, Some(next_reached), min_dist_to_end)
                {
                    return true;
                }
            }

            prev_reached = Some(next_reached);
            last_offset = Some(offset2);
        }

        let next_reached = self.reached_offsets
            .range(RangeFrom { start: target_offset_max.saturating_add(&O::one()) })
            .next();

        if last_offset.is_none()
            && self.can_improve_at_offset(bubble.bubble_exit, target_offset_min, current_score, prev_reached, next_reached, min_dist_to_end)
        {
            return true;
        }


        if last_offset.map_or(true, |v| v < target_offset_max)
            && self.can_improve_at_offset(bubble.bubble_exit, target_offset_max, current_score, prev_reached, next_reached, min_dist_to_end)
        {
            return true;
        }

        if let Some(prev) = prev_reached {
            if aln_state == AlignState::Insertion {
                let prev_reached_cost = self.visited.get_score(&AlignmentGraphNode::new(
                    bubble.bubble_exit,
                    *prev
                ), AlignState::Match);

                if prev_reached_cost + self.visited.get_costs().gap_open() > *current_score {
                    return true
                }
            } else if aln_state == AlignState::Insertion2 {
                let prev_reached_cost = self.visited.get_score(&AlignmentGraphNode::new(
                    bubble.bubble_exit,
                    *prev
                ), AlignState::Match);

                if prev_reached_cost + self.visited.get_costs().gap_open2() > *current_score {
                    return true
                }
            }
        }

        // eprintln!("PRUNE");
        false
    }

    fn can_improve_at_offset(
        &self,
        bubble_node: N,
        offset_to_check: O,
        score: &Score,
        reached_bubble_left: Option<&O>,
        reached_bubble_right: Option<&O>,
        min_dist_to_end: usize,
    ) -> bool {
        let implicit_score = match (reached_bubble_left, reached_bubble_right) {
            (None, None) => None,
            (Some(offset_left), Some(offset_right)) => {
                let left_score = self.visited.get_score(&AlignmentGraphNode::new(bubble_node, *offset_left), AlignState::Match);
                let right_score = self.visited.get_score(&AlignmentGraphNode::new(bubble_node, *offset_right), AlignState::Match);

                let gap_from_left = offset_to_check - *offset_left;
                let gap_from_right = *offset_right - offset_to_check;

                let implicit_score_from_left = left_score
                    + self.visited.get_costs().gap_cost(AlignState::Match, gap_from_left.as_usize());

                let implicit_score_from_right = right_score
                    + self.visited.get_costs().gap_cost(AlignState::Match, gap_from_right.as_usize());

                // eprintln!("Bubbles visited ({offset_left:?}, {offset_right:?}), scores: ({implicit_score_from_left:?}, {implicit_score_from_right:?})");

                // Only take deletion score into consideration if the deletion length is smaller than
                // the path length to the POA graph end node
                if gap_from_right.as_usize() > min_dist_to_end {
                    Some(implicit_score_from_left)
                } else {
                    Some(min(implicit_score_from_left, implicit_score_from_right))
                }
            },
            (None, Some(offset_right)) => {
                let right_score = self.visited.get_score(&AlignmentGraphNode::new(bubble_node, *offset_right), AlignState::Match);
                let gap_from_right = *offset_right - offset_to_check;
                let implicit_score_from_right = right_score
                    + self.visited.get_costs().gap_cost(AlignState::Match, gap_from_right.as_usize());

                // eprintln!("Bubbles visited (None, {offset_right:?}), scores: (None, {implicit_score_from_right:?})");

                // Only take deletion score into consideration if the deletion length is smaller than
                // the path length to the POA graph end node
                if gap_from_right.as_usize() > min_dist_to_end {
                    None
                } else {
                    Some(implicit_score_from_right)
                }
            },
            (Some(offset_left), None) => {
                let left_score = self.visited.get_score(&AlignmentGraphNode::new(bubble_node, *offset_left), AlignState::Match);
                let gap_from_left = offset_to_check - *offset_left;
                let implicit_score_from_left = left_score
                    + self.visited.get_costs().gap_cost(AlignState::Match, gap_from_left.as_usize());

                // eprintln!("Bubbles visited ({offset_left:?}, None), scores: ({implicit_score_from_left:?}, None)");

                Some(implicit_score_from_left)
            }
        };

        // eprintln!("Implicit gap score: {implicit_score:?}, score from align state: {score:?}");
        implicit_score.map_or(true, |s| *score < s)
    }
}