def main():
parser = argparse.ArgumentParser(description = "pairwise local alignment")
parser.add_argument("gap_cost", help="affine gap penalty cost", type = int)
parser.add_argument("gap_extension", help = "gap extension cost", type = int)
parser.add_argument("matrix", help="scoring matrix")
parser.add_argument("--score_only", help="set True if you only want the alignment score, False otherwise", type=str2bool, nargs="?", const=True, default=True)
args = parser.parse_args()
matrix_path = os.path.join("/Users/matt/OneDrive/UCSF/algorithms/HW3/scoring_matrices/", args.matrix)
with open("/Users/matt/OneDrive/UCSF/algorithms/HW3/metadata/Negpairs.txt") as tsvfile:
reader = csv.reader(tsvfile, delimiter = " ")
for x, pair in enumerate(reader):
A_input = os.path.join("/Users/matt/OneDrive/UCSF/algorithms/HW3/", pair[0])
B_input = os.path.join("/Users/matt/OneDrive/UCSF/algorithms/HW3/", pair[1])
fasta_seq = SeqIO.parse(open(A_input),"fasta")
seqarrayA = np.empty((50),dtype=np.str)
seqarrayB = np.empty((50),dtype=np.str)
print(seqarrayA)
for fasta in fasta_seq:
seq_A = str(fasta.seq).upper()
seqarrayA[x] = seq_A
fasta_seqb = SeqIO.parse(open(B_input),"fasta")
for fasta in fasta_seqb:
seq_B = str(fasta.seq).upper()
seqarrayB[x] = seq_B
print(seq_A)
fp_matrix=np.zeros((20,5))
for i in range(0,20,5):
for j in range(0,5,2):
with open("/Users/matt/OneDrive/UCSF/algorithms/HW3/metadata/Negpairs.txt") as tsvfile:
reader = csv.reader(tsvfile, delimiter=" ")
scores_array = np.zeros((50))
for x, pair in enumerate(reader):
#Loop through pairs
A_input = os.path.join("/Users/matt/OneDrive/UCSF/algorithms/HW3/", pair[0])
B_input = os.path.join("/Users/matt/OneDrive/UCSF/algorithms/HW3/", pair[1])
fasta_seq = SeqIO.parse(open(A_input),"fasta")
for fasta in fasta_seq:
seq_A = str(fasta.seq).upper()
fasta_seqb= SeqIO.parse(open(B_input), "fasta")
for fasta in fasta_seqb:
seq_B = str(fasta.seq).upper()
M = scoringMatrixParse(matrix_path)
H = matrix(seq_A, seq_B, M, -(i+1), -(j+1))
s = traceback(H, args.score_only, b=seq_B, b_="", old_i=0)
scores_array[x] = s
fpr = sum(scores_array>311.8)/50
fp_matrix[i,j] = fpr
print(fp_matrix)
def test_matrix():
############-- Online check --############################
# Program: water
# Rundate: Mon 25 Feb 2019 00:21:17
# Commandline: water
# -auto
# -stdout
# -asequence emboss_water-I20190225-002115-0388-17773373-p2m.asequence
# -bsequence emboss_water-I20190225-002115-0388-17773373-p2m.bsequence
# -datafile EBLOSUM50
# -gapopen 10.0
# -gapextend 1.0
# -aformat3 pair
# -sprotein1
# -sprotein2
# Align_format: pair
# Report_file: stdout
########################################
#=======================================
#
# Aligned_sequences: 2
# 1: HBA_HUMAN
# 2: HBA_MOUSE
# Matrix: EBLOSUM50
# Gap_penalty: 10.0
# Extend_penalty: 1.0
#
# Length: 26
# Identity: 18/26 (69.2%)
# Similarity: 21/26 (80.8%)
# Gaps: 0/26 ( 0.0%)
# Score: 135.0
#
#
#=======================================
seqA = "MVLSPADKTNVKAAWGKVGAHAGEYG"
seqB = "MVLSGEDKSNIKAAWGKIGGHGAEYGAE"
M = matrix(seqA, seqB, B50, -10, -1)
s = traceback(M, True, b=seqB, b_="", old_i=0)
assert s == 135 #135 should be the score (shown above)
def test_alignment_format():
seqA = "AtCtggTTcc"
seqB = "atcTgccTcT"
M = matrix(seqA, seqB, B50, -9, -3)
assert np.issubdtype(M.dtype, np.dtype(
'int64')) #assert a matrix is produced dispite lowercase amino acids
negative = np.array(list(zip(nega_list, negb_list))[:4])
positive = np.array(list(zip(posa_list, posb_list))[:4])
optimized_matrix = mutate(negative, positive)
pd.DataFrame(optimized_matrix)
optimized_matrix.shape
np.reshape(optimized_matrix, (22,22)).shape
s_list = np.zeros((50))
p_list = np.zeros((50))
fpr_list = []
roc_list = []
M = scoringMatrixParse(os.path.join("/Users/matt/OneDrive/UCSF/algorithms/HW3/scoring_matrices/", "PAM250"))
M.shape
M = optimized_matrix
for x, a in enumerate(list(zip(nega_list,negb_list))[:1]):
H = matrix(a[0], a[1], M, -9, -3)
pd.DataFrame(H)
s = traceback(H, True, b=seq_B, b_="", old_i=0)
s_list[x] = s
for x, a in enumerate(zip(posa_list,posb_list)):
H = matrix(a[0], a[1], M, -9, -3)
s = traceback(H, True, b=seq_B, b_="", old_i=0)
p_list[x] = s
all_values = set(np.append(p_list, s_list).flatten())
tpr_list = []
fpr_list = []
for value in all_values:
tpr = sum(p_list > value)/len(p_list)
tpr_list.append(tpr)
fpr = sum(s_list > value)/len(s_list)