Fix the assignment of multiple quantification window coordinates (#38) (

pinellolab#403) * Mckay/pd warnings (#45) * refactor errors='ignore' to try except * refactored integer slice to iloc[] * moved to_numeric try except to function * Refactor to_numeric_ignore_errors to to_numeric_ignore_columns This change is slightly cleaner because it addresses the root issue that some columns are strings (and can therefore not be converted to numeric types). Now if an error does occur when converting the dfs to numeric types it won't be swallowed up. * Add documentation to to_numeric_ignore_columns --------- * Extract out quantification window coordinate function * Refactor get_quant_window_coordinates function into two The rationale behind this is that the behavior around the cloned amplicon is quite different than if the qwc are specified directly for the amplicon. * Handling qwc: add unit tests, refactor some more and add documentation * Extract out get_relative_coordinates function This function just computes the relative indexes without doing an alignment. * Add clarifying unit tests for `get_relative_coordinates` * Refactor cloned indexes to use ref_positions instead of s1inds * fixed function for getting cloned qwc idxs * added tests for cloned qwc function * Introduce pandas sorting in CRISPRessoCompare (#47) * Fix interleaved fastq input in CRISPRessoPooled and suppress CRISPRessoWGS params (#42) * Extract out split_interleaved_fastq function to CRISPRessoShared * Implement splitting interleaved fastq files in CRISPRessoPooled * Suppress split_interleaved_input from CRISPRessoWGS parameters * Suppress other parameters in CRISPRessoWGS * Move where interleaved fastq files are split to be trimmed properly * Bug Fix - 367 (#35) * - Fixed references to ref_names_for_pe * removed extra tabs * trying to match empty line, no tabs * - changed references to ref_names[0] * Mckay/pd warnings (#45) * refactor errors='ignore' to try except * refactored integer slice to iloc[] * moved to_numeric try except to function * Refactor to_numeric_ignore_errors to to_numeric_ignore_columns This change is slightly cleaner because it addresses the root issue that some columns are strings (and can therefore not be converted to numeric types). Now if an error does occur when converting the dfs to numeric types it won't be swallowed up. * Add documentation to to_numeric_ignore_columns --------- --------- * removed if check * implemented last test * changed NT to BadParameterException * changed tests, NT to BadParameter exceptions * Uncomment and correct tests for `get_relative_coordinates` * finished qwc tests * 0 is an acceptable qwc * new get_relative_coords function * added relative coordinate tests * removed unused functions * formatting * check for 0 qwc * remove test code * remove comment * Move read filtering to after merging in CRISPResso (#39) * Move read filtering to after merging This is in an effort to be consistent with the behavior and results of CRISPRessoPooled. * Properly assign the correct file names for read filtering * Add space around operators * GitHub actions on pr (#51) * Run integration tests on pull_request * Run pytest on pull_request * Run pylint on pull_request * Run tests on PR only when opening PR (#53) * Update reports (#52) * Update report changes * Switch branch of integration test repo * Remove extraneous `crispresso_data_path` * Point integration tests back to master --------- Co-authored-by: mbowcut2 <[email protected]> Co-authored-by: McKay <[email protected]> Co-authored-by: Samuel Nichols <[email protected]>
edilytics · Nov 8, 2024 · 7763265 · 7763265
1 parent b772568
commit 7763265
Show file tree

Hide file tree

Showing 2 changed files with 49 additions and 15 deletions.
diff --git a/CRISPResso2/CRISPRessoCORE.py b/CRISPResso2/CRISPRessoCORE.py
@@ -232,7 +232,7 @@ def get_cloned_include_idxs_from_quant_window_coordinates(quant_window_coordinat
             idxs[i] = -1 * abs(idxs[i])
     for coord in split_quant_window_coordinates(quant_window_coordinates):
         include_idxs.extend(idxs[coord[0]:coord[1] + 1])
-
+        
     return list(filter(lambda x: x >= 0, include_idxs))
 
 
@@ -2205,21 +2205,15 @@ def get_prime_editing_guides(this_amp_seq, this_amp_name, ref0_seq, prime_edited
                     refs[ref_name]['contains_guide'] = refs[clone_ref_name]['contains_guide']
 
                 #quantification window coordinates override other options
-                if amplicon_quant_window_coordinates_arr[clone_ref_idx] != "":
+                if amplicon_quant_window_coordinates_arr[clone_ref_idx] != "" and amplicon_quant_window_coordinates_arr[this_ref_idx] != '0':
                     if amplicon_quant_window_coordinates_arr[this_ref_idx] != "":
-                        this_quant_window_coordinates = amplicon_quant_window_coordinates_arr[this_ref_idx]
+                        this_include_idxs = get_include_idxs_from_quant_window_coordinates(amplicon_quant_window_coordinates_arr[this_ref_idx])
                     else:
-                        this_quant_window_coordinates = amplicon_quant_window_coordinates_arr[clone_ref_idx]
-                    this_include_idxs = []
-                    these_coords = this_quant_window_coordinates.split("_")
-                    for coord in these_coords:
-                        coordRE = re.match(r'^(\d+)-(\d+)$', coord)
-                        if coordRE:
-                            start = s1inds[int(coordRE.group(1))]
-                            end = s1inds[int(coordRE.group(2)) + 1]
-                            this_include_idxs.extend(range(start, end))
-                        else:
-                            raise NTException("Cannot parse analysis window coordinate '" + str(coord))
+                        this_include_idxs = get_cloned_include_idxs_from_quant_window_coordinates(
+                            amplicon_quant_window_coordinates_arr[clone_ref_idx],
+                            s1inds.copy(),
+                        )
+
                     #subtract any indices in 'exclude_idxs' -- e.g. in case some of the cloned include_idxs were near the read ends (excluded)
                     this_exclude_idxs = sorted(list(set(refs[ref_name]['exclude_idxs'])))
                     this_include_idxs = sorted(list(set(np.setdiff1d(this_include_idxs, this_exclude_idxs))))

diff --git a/CRISPResso2/CRISPRessoShared.py b/CRISPResso2/CRISPRessoShared.py
@@ -1732,7 +1732,47 @@ def set_guide_array(vals, guides, property_name):
     for idx, val in enumerate(vals_array):
         if val != '':
             ret_array[idx] = int(val)
-    return ret_array
+    return ret_array        
+
+
+def get_relative_coordinates(to_sequence, from_sequence):
+    """Given an alignment, get the relative coordinates of the second sequence to the first.
+
+    For example, from_sequence[i] matches to to_sequence[inds[i]]. A `-1`
+    indicates a gap at the beginning of `to_sequence`.
+
+    Parameters
+    ----------
+    to_sequence : str
+        The alignment of the first sequence (where the coordinates are relative to)
+    from_sequence : str
+        The alignment of the second sequence
+
+    Returns
+    -------
+    s1inds_gap_left : list of int
+        The relative coordinates of the second sequence to the first, where gaps
+        in the first sequence are filled with the left value.
+    s1inds_gap_right : list of int
+        The relative coordinates of the second sequence to the first, where gaps
+        in the first sequence are filled with the right value.
+    """
+    s1inds_gap_left = []
+    s1inds_gap_right = []
+    s1idx_left = -1
+    s1idx_right = 0
+    s2idx = -1
+    for ix in range(len(to_sequence)):
+        if to_sequence[ix] != "-":
+            s1idx_left += 1
+        if from_sequence[ix] != "-":
+            s2idx += 1
+            s1inds_gap_left.append(s1idx_left)
+            s1inds_gap_right.append(s1idx_right)
+        if to_sequence[ix] != "-":
+            s1idx_right += 1
+
+    return s1inds_gap_left, s1inds_gap_right
 
 
 def get_relative_coordinates(to_sequence, from_sequence):