diff --git a/crisprbact/predict.py b/crisprbact/predict.py
index 20e97ddbe977b7aff70b432d9dd3247319b95e5a..00f3aeabcba89cb2d18245d2620d84d21aa270bf 100644
--- a/crisprbact/predict.py
+++ b/crisprbact/predict.py
@@ -43,7 +43,12 @@ def find_targets(seq):
)
-def on_target_predict(seq, genome=None, seed_size=7):
+def get_strand_value(value):
+ strand_dict = {"+": 1, "1": 1, "-": -1, "-1": -1}
+ return strand_dict[str(value)]
+
+
+def on_target_predict(seq, genome=None, seed_sizes=[8, 9, 10, 11, 12]):
seq = seq.upper() # make uppercase
seq = re.sub(r"\s", "", seq) # removes white space
@@ -65,47 +70,80 @@ def on_target_predict(seq, genome=None, seed_size=7):
preds = predict(X)
for i, target in enumerate(alltargets):
+ target_id = i + 1
+ target.update({"id": target_id})
target.update({"pred": preds[i]})
- target.update({"seed_size": seed_size})
if genome:
- off_target_df = compute_off_target_df(
- target["guide"], seed_size, records, genome_features
- )
- off_targets_list = []
- if not off_target_df.empty:
- off_targets = off_target_df.loc[
- 0:, ["start", "end", "pampos", "strand", "recid", "features"]
- ]
- for index, off_t in enumerate(off_targets.values.tolist()):
- off_target_dict = {
- "off_target_start": off_t[0],
- "off_target_end": off_t[1],
- "pampos": off_t[2],
- "strand": off_t[3],
- "recid": off_t[4],
- }
- if len(off_t[5]) > 0:
- # Loop for each feature
- for feat in off_t[5]:
- feature_dict = {
- "feat_strand": feat.location.strand,
- "feat_start": feat.location.start,
- "feat_end": feat.location.end,
- "feat_type": feat.type,
- }
- for k, feat in feat.qualifiers.items():
- if k != "translation":
- feature_dict[k] = "::".join(feat)
- off_targets_list.append(
- {**feature_dict, **off_target_dict}
+ off_targets_per_seed = []
+ for seed_size in seed_sizes:
+ off_target_df = compute_off_target_df(
+ target["guide"], seed_size, records, genome_features
+ )
+ off_targets_list = []
+ if not off_target_df.empty:
+ off_targets = off_target_df.loc[
+ 0:,
+ ["start", "end", "pampos", "strand", "recid", "features"],
+ ]
+ for j, off_t in enumerate(off_targets.values.tolist()):
+ off_target_dict = {
+ "id": str(target_id)
+ + "-"
+ + str(seed_size)
+ + "-"
+ + str(j),
+ "off_target_start": off_t[0],
+ "off_target_end": off_t[1],
+ "off_target_pampos": off_t[2],
+ "off_target_strand": off_t[3],
+ "off_target_recid": off_t[4],
+ }
+
+ off_t[5] = list(
+ filter(
+ lambda feat: get_strand_value(
+ off_target_dict["off_target_strand"]
+ )
+ != get_strand_value(feat.location.strand),
+ off_t[5],
)
- else:
- off_targets_list.append(off_target_dict)
- target.update({"off_targets": off_targets_list})
- else:
- target.update({"off_targets": off_targets_list})
+ )
+ if len(off_t[5]) > 0:
+ # Loop for each feature
+ for feat in off_t[5]:
+ feature_dict = {
+ "off_target_feat_strand": feat.location.strand,
+ "off_target_feat_start": feat.location.start,
+ "off_target_feat_end": feat.location.end,
+ "off_target_feat_type": feat.type,
+ }
+ for k, feat in feat.qualifiers.items():
+ if k != "translation":
+ feature_dict[k] = "::".join(feat)
+ off_targets_list.append(
+ {**feature_dict, **off_target_dict}
+ )
+ else:
+ off_targets_list.append(off_target_dict)
+ off_targets_per_seed.append(
+ {
+ "id": str(i) + "-" + str(seed_size),
+ "seed_size": seed_size,
+ "off_targets": off_targets_list,
+ }
+ )
+ else:
+ off_targets_per_seed.append(
+ {
+ "id": str(i) + "-" + str(seed_size),
+ "seed_size": seed_size,
+ "off_targets": off_targets_list,
+ }
+ )
+ # target.update({"off_targets": off_targets_list})
+ target.update({"off_targets_per_seed": off_targets_per_seed})
else:
- target.update({"off_targets": []})
+ target.update({"off_targets_per_seed": []})
return alltargets
else:
return []