v1.0.1 (#18)

* Fix FID+IID filtering, simplify based on tuples.

* Delete unused class

* Delete unused class 2

* Bump version

* Extra comments for clarity

* bump patch version instead

---------

Co-authored-by: Benjamin Wingfield <[email protected]>

pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "fraposa-pgsc"
-version = "1.0.0"
+version = "1.0.1"
 description = "Tools to perform ancestry projection to a reference dataset within the calculator pipeline (pgsc_calc)"
 homepage = "https://github.com/PGScatalog/fraposa_pgsc"
 authors = ["smlmbrt <[email protected]>"]

src/fraposa_pgsc/fraposa.py

@@ -22,7 +22,6 @@
 from sklearn.utils.extmath import randomized_svd
 from typing import Union
 
-from .sampleid import SampleID
 
 def create_logger(out_filepref='fraposa'):
     log = logging.getLogger()
@@ -139,25 +138,24 @@ def read_bed(bed_filepref, dtype=np.int8, filt_iid=None):
     n = len(fam)
 
     if filt_iid:
-        fam_ids = set(SampleID(x, y) for x, y in zip(fam['fid'], fam['iid'], strict=True))
-        matched_ids = filt_iid.intersection(fam_ids)
-        if len(matched_ids) == 0:
+        fam_ids = list(zip(fam['fid'], fam['iid'])) # create tuples of ids from genotyping files
+        fam_mask = [x in filt_iid for x in fam_ids] # T/F overlap of genotype data and filter IDs (tuples)
+        n_matched = sum(fam_mask)
+        if n_matched == 0:
             raise ValueError(f"ERROR: 0 / {len(filt_iid)} ids in filter list match the study dataset")
-        elif len(fam_ids) != len(fam):
+        elif len(set(fam_ids)) != len(fam):
             raise ValueError("Samples with duplicated FID + IID detected, please remove and retry")
 
-        bed = np.zeros(shape=(p, len(matched_ids)), dtype=dtype)
-        # in will call SampleID's __hash__ method which uses (fid, iid)
-        fam_mask = pd.Series((x in matched_ids for x in fam_ids), dtype=bool)
-        i_extract = np.where(fam_mask == True)
+        bed = np.zeros(shape=(p, n_matched), dtype=dtype)
+        i_extract = [i for i,x in enumerate(fam_mask) if x is True] # idx to extract from genotype matrix
         for (i, (snp, genotypes)) in enumerate(pyp):
             bed[i,:] = genotypes[i_extract]
         fam = fam.loc[fam_mask,:]
-        if len(matched_ids) < len(filt_iid):
-            logging.warning('Warning: only {} / {} ids in filter list match the study dataset'.format(len(matched_ids),
+        if n_matched < len(filt_iid):
+            logging.warning('Warning: only {} / {} ids in filter list match the study dataset'.format(n_matched,
                                                                                                       len(filt_iid)))
         else:
-            logging.info('Extracted {} samples from study genotyping data'.format(len(matched_ids)))
+            logging.info('Extracted {} samples from study genotyping data'.format(n_matched))
     else:
         bed = np.zeros(shape=(p, n), dtype=dtype)
         for (i, (snp, genotypes)) in enumerate(pyp):

src/fraposa_pgsc/fraposa_runner.py

@@ -2,7 +2,6 @@
 import csv
 
 import fraposa_pgsc.fraposa as fp
-from fraposa_pgsc.sampleid import SampleID
 import argparse
 
 
@@ -38,8 +37,17 @@ def main():
 
     try:
         with open(args.stu_filt_iid) as f:
-            reader = csv.reader(f, delimiter="\t")
-            stu_filt_iid = set(SampleID(x[0], x[1]) for x in list(reader))
+            reader = csv.reader(f, delimiter="\t") # reads columns as str
+            stu_filt_iid = []
+            for x in reader:
+                if x[0] == "0":
+                    stu_filt_iid.append((x[1],x[1])) # replace missing FID with IID
+                else:
+                    stu_filt_iid.append((x[0], x[1])) # return FID, IID
+            l_input = len(stu_filt_iid)
+            stu_filt_iid = set(stu_filt_iid)
+            if l_input != len(stu_filt_iid):
+                raise ValueError("Duplicate IDs found in filter list")
     except TypeError:
         stu_filt_iid = None
     except IndexError: