def get_linreg_optimal_codons(linreg_dir, aa_seq, maximum=False):
linreg_data_fname = linreg_dir + "/linreg_data.txt"
d = proc.load_linreg_data_file(linreg_data_fname)
rel_cod_idxs = d["rel_cod_idxs"]
wts = proc.load_obj(linreg_dir + "/wts.pkl")
opt_total_seq, opt_vit_score = opt.get_optimal_codons_linreg(
aa_seq, wts, rel_cod_idxs, maximum=maximum)
return opt_total_seq, opt_vit_score
def get_protein_score_dist(nn_dir, epoch, aa_seq, num_samples, nt_feats=False):
init_data_pkl = nn_dir + "/init_data/init_data.pkl"
params = proc.load_obj(init_data_pkl)
rel_cod_idxs = params["rel_cod_idxs"]
my_nn = load_lasagne_feedforward_nn(nn_dir, epoch)
scores_sorted, cod_seqs_sorted = opt.get_score_dist(aa_seq,
my_nn,
rel_cod_idxs,
num_samples,
nt_feats=nt_feats)
return scores_sorted, cod_seqs_sorted
def load_lasagne_adjacency_nn(nn_dir, epoch):
init_data_pkl = nn_dir + "/init_data/init_data.pkl"
params = proc.load_obj(init_data_pkl)
try:
max_struc_start_idx = params["max_struc_start_idx"],
except KeyError:
max_struc_start_idx = None
try:
max_struc_width = params["max_struc_width"],
except KeyError:
max_struc_width = None
X_tr, y_tr, X_te, y_te = proc.load_lasagne_data(
params["gene_len_fname"],
params["gene_seq_fname"],
params["tr_codons_fname"],
params["te_codons_fname"],
params["outputs_fname"],
rel_cod_idxs=params["rel_cod_idxs"],
rel_nt_idxs=params["rel_nt_idxs"],
rel_struc_idxs=params["rel_struc_idxs"],
struc_fname=params["struc_fname"],
max_struc_start_idx=max_struc_start_idx,
max_struc_width=max_struc_width,
filter_pct=params["filter_pct"])
my_nn = lasagnenn.AdjacencyMLP(X_tr,
y_tr,
X_te,
y_te,
name=params["name"],
out_dir=params["out_dir"],
rel_cod_idxs=params["rel_cod_idxs"],
cod_adj_idxs=params["cod_adj_idxs"],
rel_nt_idxs=params["rel_nt_idxs"],
nt_adj_idxs=params["nt_adj_idxs"],
learning_rate=params["learning_rate"],
update_method=params["update_method"],
widths=params["widths"],
nonlinearity=params["nonlinearity"],
input_drop_rate=params["input_drop_rate"],
hidden_drop_rate=params["hidden_drop_rate"],
num_outputs=params["num_outputs"],
momentum=params["momentum"],
batch_size=params["batch_size"],
reloaded=True)
my_nn.unpickle_epoch(epoch)
return my_nn
def load_lasagne_feedforward_nn(nn_dir, epoch):
init_data_pkl = nn_dir + "/init_data/init_data.pkl"
params = proc.load_obj(init_data_pkl)
#Kludge, take this out in the future
if not params.get("max_struc_start_idx", False):
params["max_struc_start_idx"] = None
if not params.get("max_struc_width", False):
params["max_struc_width"] = None
if not params.get("aa_feats", False):
params["aa_feats"] = False
if not params.get("nonnegative", False):
params["nonnegative"] = False
X_tr, _, X_te, _ = proc.load_lasagne_data(
params["gene_len_fname"],
params["gene_seq_fname"],
params["tr_codons_fname"],
params["te_codons_fname"],
params["outputs_fname"],
rel_cod_idxs=params["rel_cod_idxs"],
rel_nt_idxs=params["rel_nt_idxs"],
rel_struc_idxs=params["rel_struc_idxs"],
struc_fname=params["struc_fname"],
max_struc_start_idx=params["max_struc_start_idx"],
max_struc_width=params["max_struc_width"],
aa_feats=params["aa_feats"],
filter_pct=params["filter_pct"])
my_nn = lasagnenn.FeedforwardMLP(
X_tr,
params["y_tr"],
X_te,
params["y_te"],
name=params["name"],
out_dir=params["out_dir"],
learning_rate=params["learning_rate"],
update_method=params["update_method"],
widths=params["widths"],
nonlinearity=params["nonlinearity"],
input_drop_rate=params["input_drop_rate"],
hidden_drop_rate=params["hidden_drop_rate"],
num_outputs=params["num_outputs"],
momentum=params["momentum"],
batch_size=params["batch_size"],
nonnegative=params["nonnegative"],
reloaded=True)
my_nn.unpickle_epoch(epoch)
return my_nn
def get_lasagne_optimal_codons(nn_dir,
epoch,
aa_seq,
nn_type="Feedforward",
maximum=False,
nt_feats=False):
init_data_pkl = nn_dir + "/init_data/init_data.pkl"
params = proc.load_obj(init_data_pkl)
if nn_type == "Feedforward":
my_nn = load_lasagne_feedforward_nn(nn_dir, epoch)
elif nn_type == "Adjacency":
my_nn = load_lasagne_adjacency_nn(nn_dir, epoch)
elif nn_type == "Split":
my_nn = load_lasagne_split_nn(nn_dir, epoch)
else:
print "nn_type must be in [Feedforward, Adjacency, Split]"
rel_cod_idxs = params["rel_cod_idxs"]
opt_total_seq, opt_vit_score = opt.get_optimal_codons_lasagne(
aa_seq, my_nn, rel_cod_idxs, maximum=maximum, nt_feats=nt_feats)
return opt_total_seq, opt_vit_score