view trips_create_annotation/create_annotation_sqlite.py @ 2:d70696d3341e draft

Uploaded

# Python3 script which takes in an annotation file(gtf/gff3) and a transcriptomic fasta file
# and produces an sqlite file which can be uploaded to Trips-Viz
# All co-ordinates produced are 1 based
# All start codon positions (including cds_start) should be at the first nucleotide of the codon
# All stop codon positions (including cds_stop) should be at the last nucleotide of the codon
import sys
import re
import sqlite3
from intervaltree import Interval, IntervalTree
import itertools
from sqlitedict import SqliteDict
import os

organism = sys.argv[1]
# This should be a GTF or GFF3 file
annotation_file = open(sys.argv[2], "r")
# This needs to be the transcriptomic fasta file
fasta_file = open(sys.argv[3], "r")
# This value will be added used to create UTRs of this length, useful when looking at transcriptomes without annotated UTRs
pseudo_utr_len = int(sys.argv[4])
# An example of a transcript_id from the annotation file, e.g ENST000000123456
user_transcript_id = sys.argv[5]
# An example of a gene name from the annotation file
user_gene_name = sys.argv[6]
# Set to true if transcript version is included in transcript_id, e.g: ENST000000123456.1
TRAN_VERSION = True


if os.path.isfile("{}.sqlite".format(organism)):
    print("{}.sqlite already exists".format(organism))
    sys.exit()


# old_exons = SqliteDict(
#     "/home/data2/www/tripsviz/tripsviz/trips_annotations/mus_musculus/transcriptomic_to_genomic.sqlite"
# )


delimiters = {
    "transcripts": {"before": [], "after": ["."], "annot_types": ["cds", "utr"]},
    "genes": {"before": [], "after": ['"'], "annot_types": ["lnc_rna"]},
}

punctuation = [";", " ", "-", ":", "-", ".", "=", "\t"]
# Find delimiters in the annotation and fasta files using the user_transcript_id
# and user_gene_name examples given by user.
for line in annotation_file:
    if user_transcript_id in line:
        tabsplitline = line.split("\t")
        annot_type = tabsplitline[2]
        if annot_type not in delimiters["transcripts"]["annot_types"]:
            delimiters["transcripts"]["annot_types"].append(annot_type.lower())
        splitline = line.split(user_transcript_id)
        before_delimiter = splitline[0]
        for item in punctuation:
            if item in before_delimiter:
                if len(before_delimiter.split(item)[-1]) >= 5:
                    before_delimiter = before_delimiter.split(item)[-1]
        after_delimiter = splitline[1][:2]
        if (
            before_delimiter not in delimiters["transcripts"]["before"]
            and len(before_delimiter) >= 5
        ):
            delimiters["transcripts"]["before"].append(before_delimiter)
        if after_delimiter not in delimiters["transcripts"]["after"]:
            delimiters["transcripts"]["after"].append(after_delimiter)
    if user_gene_name in line:
        tabsplitline = line.split("\t")
        annot_type = tabsplitline[2]
        print("ANNOT TYPE", annot_type)
        if annot_type not in delimiters["genes"]["annot_types"]:
            delimiters["genes"]["annot_types"].append(annot_type.lower())
        splitline = line.split(user_gene_name)
        before_delimiter = splitline[0]
        for item in punctuation:
            if item in before_delimiter:
                if len(before_delimiter.split(item)[-1]) >= 5:
                    before_delimiter = before_delimiter.split(item)[-1]
        after_delimiter = splitline[1][0]
        if (
            before_delimiter not in delimiters["genes"]["before"]
            and len(before_delimiter) >= 5
        ):
            delimiters["genes"]["before"].append(before_delimiter)
        if after_delimiter not in delimiters["genes"]["after"]:
            if after_delimiter in punctuation:
                delimiters["genes"]["after"].append(after_delimiter)

print("delimeters[genes]", delimiters["transcripts"]["annot_types"])

for line in fasta_file:
    if user_transcript_id in line:
        splitline = line.split(user_transcript_id)
        before_delimiter = splitline[0]
        for item in punctuation:
            if item in before_delimiter:
                if len(before_delimiter.split(item)[1]) >= 5:
                    before_delimiter = before_delimiter.split(item)[1]
        after_delimiter = splitline[1][0]
        if (
            before_delimiter not in delimiters["transcripts"]["before"]
            and len(before_delimiter) >= 5
        ):
            delimiters["transcripts"]["before"].append(before_delimiter)
        if after_delimiter not in delimiters["transcripts"]["after"]:
            delimiters["transcripts"]["after"].append(after_delimiter)
fasta_file.close()
annotation_file.close()


if delimiters["transcripts"]["before"] == []:
    print(
        "ERROR: No transcript_id with the name {} could be found in the annotation file".format(
            user_transcript_id
        )
    )
    sys.exit()
if delimiters["genes"]["before"] == []:
    print(
        "ERROR: No gene with the name {} could be found in the annotation file".format(
            user_gene_name
        )
    )
    sys.exit()

master_dict = {}
coding_dict = {}
notinfasta = open("notinfasta.csv", "w")

# Given a nucleotide sequence returns the positions of all start and stop codons.
def get_start_stops(transcript_sequence):
    transcript_sequence = transcript_sequence.upper()
    start_codons = ["ATG"]
    stop_codons = ["TAA", "TAG", "TGA"]
    seq_frames = {"starts": [], "stops": []}
    for codons, positions in ((start_codons, "starts"), (stop_codons, "stops")):
        if len(codons) > 1:
            pat = re.compile("|".join(codons))
        else:
            pat = re.compile(codons[0])
        for m in re.finditer(pat, transcript_sequence):
            # Increment position by 1, Frame 1 starts at position 1 not 0,
            # if it's a stop codon add another 2 so it points to the last nuc of the codon
            if positions == "starts":
                start = m.start() + 1
            else:
                start = m.start() + 3
            seq_frames[positions].append(start)
    return seq_frames


# parse fasta to get the nucleotide sequence of transcripts and the positions of start/stop codons.
fasta_file = open(sys.argv[3], "r")
read_fasta = fasta_file.read()
split_fasta = read_fasta.split(">")
for entry in split_fasta[1:]:
    newline_split = entry.split("\n")
    tran = newline_split[0]
    for item in delimiters["transcripts"]["after"]:
        if item in tran:
            tran = tran.split(item)[0]
    tran = tran.replace("-", "_").replace("(", "").replace(")", "")
    seq = "".join(newline_split[1:])
    if "_PAR_Y" in tran:
        tran += "_chrY"
    elif "_PAR_X" in tran:
        tran += "_chrX"
    tran = tran.upper()
    starts_stops = get_start_stops(seq)
    print("tran", tran)
    if tran not in master_dict:
        master_dict[tran] = {
            "utr": [],
            "cds": [],
            "exon": [],
            "start_codon": [],
            "stop_codon": [],
            "start_list": starts_stops["starts"],
            "stop_list": starts_stops["stops"],
            "transcript": [],
            "strand": "",
            "gene_name": "",
            "chrom": "",
            "seq": seq,
            "cds_start": "NULL",
            "cds_stop": "NULL",
            "length": len(seq),
            "principal": 0,
            "version": "NULL",
        }


def to_ranges(iterable):
    tup_list = []
    iterable = sorted(set(iterable))
    for key, group in itertools.groupby(enumerate(iterable), lambda t: t[1] - t[0]):
        group = list(group)
        tup_list.append((group[0][1], group[-1][1]))
    return tup_list


# parse annotation file to get chromsome, exon location and CDS info for each transcript
def parse_gtf_file(annotation_file):
    for line in annotation_file:
        if line == "\n":
            continue
        if line[0] != "#":
            splitline = (line.replace("\n", "")).split("\t")
            chrom = splitline[0]
            try:
                annot_type = splitline[2].lower()
            except:
                print(
                    "ERROR tried to index to second item in splitline: ",
                    splitline,
                    line,
                )
                sys.exit()
            # if annot_type not in ["cds", "utr", "exon", "transcript","five_prime_utr", "three_prime_utr","stop_codon","start_codon"]:
            # 	continue
            if (
                annot_type not in delimiters["transcripts"]["annot_types"]
                and annot_type not in delimiters["genes"]["annot_types"]
            ):
                continue
            if annot_type == "five_prime_utr" or annot_type == "three_prime_utr":
                annot_type = "utr"
            strand = splitline[6]
            if strand == "+":
                start = int(splitline[3])
                end = int(splitline[4])
            else:
                start = int(splitline[3]) + 1
                end = int(splitline[4]) + 1
            desc = splitline[8]
            tran = desc
            gene = desc
            for item in delimiters["transcripts"]["before"]:
                if item in tran:
                    tran = tran.split(item)[1]
            for item in delimiters["transcripts"]["after"]:
                if item in tran:
                    tran = tran.split(item)[0]
            if "." in tran and TRAN_VERSION == True:
                # print ("raw tran",tran)
                tran = tran.split(".")
                version = int(tran[-1].split("_")[0])
                tran = tran[0]
            else:
                version = "NULL"
            tran = tran.replace("-", "_").replace(".", "_")
            tran = tran.replace("(", "").replace(")", "")
            tran = tran.replace(" ", "").replace("\t", "")
            tran = tran.upper()
            tran = tran.replace("GENE_", "").replace("ID_", "")
            if "_PAR_Y" in desc:
                # print ("adding _PAR_Y to tran")
                tran = tran + "_PAR_Y"
                # print ("New tran ", tran)
            # if "PAR_Y" in line:
            # 	print (line)
            # 	#sys.exit()
            # print ("tran",tran,version)
            # if tran == "ENST00000316448":
            # 	print ("annot type",annot_type)
            # 	print ("appending exon to tran", start, end,line)

            gene_found = False

            if annot_type in delimiters["genes"]["annot_types"]:
                for item in delimiters["genes"]["before"]:
                    if item in gene:
                        gene_found = True
                        gene = gene.split(item)[1]
                for item in delimiters["genes"]["after"]:
                    if item in gene:
                        gene = gene.split(item)[0]
                gene = gene.replace("'", "''")
                gene = gene.replace("GENE_", "")
                gene = gene.replace("ID_", "")
                gene = gene.upper()
            if tran in master_dict:
                master_dict[tran]["strand"] = strand
                if strand == "+":
                    if annot_type in master_dict[tran]:
                        master_dict[tran][annot_type].append((start, end))
                else:
                    if annot_type in master_dict[tran]:
                        master_dict[tran][annot_type].append((start, end))
                master_dict[tran]["chrom"] = chrom
                master_dict[tran]["version"] = version
                if gene_found == True:
                    master_dict[tran]["gene_name"] = gene
            else:
                notinfasta.write("{}\n".format(tran))


annotation_file = open(sys.argv[2], "r")
parse_gtf_file(annotation_file)


# remove transcripts that were in fasta file but not in annotation_file
notinannotation = []
for tran in master_dict:
    if master_dict[tran]["chrom"] == "":
        # print ("tran {} has no chrom :(".format(tran))
        notinannotation.append(tran)
for tran in notinannotation:
    print(tran)
    del master_dict[tran]
# Dictionary to store the coding status of a gene, if any transcript of this gene is coding, the value will be True
coding_genes_dict = {}
# parse master_dict to calculate length, cds_start/stop and exon junction positions
for tran in master_dict:
    try:
        transeq = master_dict[tran]["seq"]
    except Exception as e:
        print("not in fasta", tran)
        notinfasta.write("{}\n".format(tran))
        continue
    exon_junctions = []
    total_length = len(transeq)
    three_len = 1
    five_len = 1
    strand = master_dict[tran]["strand"]
    if master_dict[tran]["gene_name"] == "":
        master_dict[tran]["gene_name"] = tran
    gene = master_dict[tran]["gene_name"]
    if gene not in coding_genes_dict:
        coding_genes_dict[gene] = False

    if master_dict[tran]["cds"] == []:
        tran_type = "noncoding"
        cds_start = "NULL"
        cds_stop = "NULL"
    else:
        # get utr lengths from annotation
        tran_type = "coding"
        coding_genes_dict[gene] = True
        sorted_exons = sorted(master_dict[tran]["exon"])
        sorted_cds = sorted(master_dict[tran]["cds"])

        min_cds = sorted_cds[0][0]
        # Some annotation files do not have utr annotation types, so fix that here if thats the case
        if master_dict[tran]["utr"] == []:
            for exon_tup in master_dict[tran]["exon"]:
                if exon_tup not in master_dict[tran]["cds"]:
                    # Now check if this overlaps with any of the CDS exons
                    overlap = False
                    for cds_tup in master_dict[tran]["cds"]:
                        if exon_tup[0] == cds_tup[0] and exon_tup[1] != cds_tup[1]:
                            master_dict[tran]["utr"].append((cds_tup[1], exon_tup[1]))
                            overlap = True
                        if exon_tup[0] != cds_tup[0] and exon_tup[1] == cds_tup[1]:
                            master_dict[tran]["utr"].append((exon_tup[0], cds_tup[0]))
                            overlap = True
                    if overlap == False:
                        master_dict[tran]["utr"].append(exon_tup)

        for tup in sorted(master_dict[tran]["utr"]):
            if tup[0] < min_cds:
                five_len += (tup[1] - tup[0]) + 1
            elif tup[0] > min_cds:
                three_len += (tup[1] - tup[0]) + 1
            else:
                pass
        if strand == "+":
            if len(sorted_exons) > 1:
                sorted_exons[0] = (
                    sorted_exons[0][0] - pseudo_utr_len,
                    sorted_exons[0][1],
                )
                sorted_exons[-1] = (
                    sorted_exons[-1][0],
                    sorted_exons[-1][1] + pseudo_utr_len,
                )
            else:
                sorted_exons[0] = (
                    sorted_exons[0][0] - pseudo_utr_len,
                    sorted_exons[0][1] + pseudo_utr_len,
                )
            master_dict[tran]["exon"] = sorted_exons
            cds_start = five_len + pseudo_utr_len
            cds_stop = ((total_length - three_len) - pseudo_utr_len) + 4
        elif strand == "-":
            if len(sorted_exons) > 1:
                sorted_exons[0] = (
                    (sorted_exons[0][0] - pseudo_utr_len),
                    sorted_exons[0][1],
                )
                sorted_exons[-1] = (
                    sorted_exons[-1][0],
                    (sorted_exons[-1][1] + pseudo_utr_len),
                )
            else:
                sorted_exons[0] = (
                    (sorted_exons[0][0] - pseudo_utr_len),
                    (sorted_exons[0][1] + pseudo_utr_len),
                )
            master_dict[tran]["exon"] = sorted_exons
            cds_start = three_len + pseudo_utr_len
            cds_stop = ((total_length - (five_len)) - pseudo_utr_len) + 4
            # if tran == "ENST00000381401":
            # 	print ("cds start, cds stop, five_len, three_len",cds_start,cds_stop,five_len,three_len)
            # 	#sys.exit()
        else:
            print("strand is unknown: {}".format(strand))
            sys.exit()
    # get exon junctions, cds is easy just get end of each tuple except last, same for utr except for if same as cds start/stop
    total_intronic = 0
    try:
        if strand == "+":
            tx_start = min(sorted(master_dict[tran]["exon"]))[0]
            prev_end = tx_start
            for tup in sorted(master_dict[tran]["exon"])[:-1]:
                total_intronic += tup[0] - prev_end
                exon_junctions.append(((tup[1]) - tx_start) - total_intronic)
                prev_end = tup[1]
        elif strand == "-":
            tx_start = max(sorted(master_dict[tran]["exon"]))[-1]
            prev_end = tx_start
            for tup in (sorted(master_dict[tran]["exon"])[1:])[::-1]:
                total_intronic += (tup[0] + 1) - prev_end
                exon_junctions.append(((tup[1]) - tx_start) - total_intronic)
                prev_end = tup[1]
    except:
        if strand == "+":
            tx_start = min(sorted(master_dict[tran]["cds"]))[0]
            prev_end = tx_start
            for tup in sorted(master_dict[tran]["cds"])[:-1]:
                total_intronic += tup[0] - prev_end
                exon_junctions.append(((tup[1]) - tx_start) - total_intronic)
                prev_end = tup[1]
        elif strand == "-":
            tx_start = max(sorted(master_dict[tran]["cds"]))[-1]
            prev_end = tx_start
            for tup in (sorted(master_dict[tran]["cds"])[1:])[::-1]:
                total_intronic += (tup[0] + 1) - prev_end
                exon_junctions.append(((tup[1]) - tx_start) - total_intronic)
                prev_end = tup[1]
    # This can happen when a coding transcript doesn't have a properly annotated 3' trailer
    if cds_stop != "NULL":
        if cds_stop > total_length:
            cds_stop = total_length
    if strand == "+" and cds_start != "NULL":
        master_dict[tran]["cds_start"] = cds_start
        master_dict[tran]["cds_stop"] = cds_stop
    elif strand == "-" and cds_start != "NULL":
        master_dict[tran]["cds_start"] = cds_start
        master_dict[tran]["cds_stop"] = cds_stop

    master_dict[tran]["strand"] = strand
    master_dict[tran]["tran_type"] = tran_type
    master_dict[tran]["exon_junctions"] = exon_junctions

longest_tran_dict = {}
for tran in master_dict:
    try:
        gene = master_dict[tran]["gene_name"]
    except:
        continue
    if coding_genes_dict[gene] == True:
        if "cds_start" in master_dict[tran]:
            if (
                master_dict[tran]["cds_stop"] != "NULL"
                and master_dict[tran]["cds_start"] != "NULL"
            ):
                cds_length = (
                    master_dict[tran]["cds_stop"] - master_dict[tran]["cds_start"]
                )
                if gene not in longest_tran_dict:
                    longest_tran_dict[gene] = {"tran": tran, "length": cds_length}
                else:
                    if cds_length > longest_tran_dict[gene]["length"]:
                        longest_tran_dict[gene] = {"tran": tran, "length": cds_length}
                    if cds_length == longest_tran_dict[gene]["length"]:
                        if (
                            master_dict[tran]["length"]
                            > master_dict[longest_tran_dict[gene]["tran"]]["length"]
                        ):
                            longest_tran_dict[gene] = {
                                "tran": tran,
                                "length": cds_length,
                            }
    else:
        length = master_dict[tran]["length"]
        if gene not in longest_tran_dict:
            longest_tran_dict[gene] = {"tran": tran, "length": length}
        elif length > longest_tran_dict[gene]["length"]:
            longest_tran_dict[gene] = {"tran": tran, "length": length}


for gene in longest_tran_dict:
    longest_tran = longest_tran_dict[gene]["tran"]
    master_dict[longest_tran]["principal"] = 1

gene_sample = []
for key in list(master_dict)[:10]:
    try:
        gene_sample.append(master_dict[key]["gene_name"])
    except:
        pass
print(master_dict)
print("Here is a sample of the transcript ids: {}".format(list(master_dict)[:10]))
print("Here is a sample of the gene names: {}".format(gene_sample))


# Takes a transcript, transcriptomic position and a master_dict (see ribopipe scripts) and returns the genomic position, positions should be passed 1 at a time.
def tran_to_genome(tran, start_pos, end_pos, master_dict):
    pos_list = []
    for i in range(start_pos, end_pos + 1):
        pos_list.append(i)
    genomic_pos_list = []
    if tran in master_dict:
        transcript_info = master_dict[tran]
    else:
        return ("Null", [])

    chrom = transcript_info["chrom"]
    strand = transcript_info["strand"]
    exons = sorted(transcript_info["exon"])
    # print ("chrom,strand,exons",chrom,strand,exons)
    for pos in pos_list:
        # print ("pos",pos)
        if strand == "+":
            exon_start = 0
            for tup in exons:
                # print ("tup",tup)
                exon_start = tup[0]
                exonlen = tup[1] - tup[0]
                if pos > exonlen:
                    pos = (pos - exonlen) - 1
                else:
                    break
            # print ("appending exon_start-pos", exon_start, pos, exon_start+pos)
            genomic_pos_list.append((exon_start + pos) - 1)
        elif strand == "-":
            exon_start = 0
            for tup in exons[::-1]:
                # print ("tup",tup)
                exon_start = tup[1]
                exonlen = tup[1] - tup[0]
                # print ("exonlen",exonlen)
                if pos > exonlen:
                    # print ("pos is greater")
                    pos = (pos - exonlen) - 1
                    # print ("new pos",pos)
                else:
                    break
            # print ("appending exon_start-pos", exon_start, pos, exon_start-pos)
            genomic_pos_list.append((exon_start - pos) + 1)
    return (chrom, genomic_pos_list)


orf_dict = {
    "uorf": {},
    "ouorf": {},
    "cds": {},
    "nested": {},
    "odorf": {},
    "dorf": {},
    "minusone": {},
    "readthrough": {},
    "plusone": {},
    "noncoding": {},
    "extension": {},
    "inframe_stop": {},
}

start_codons = ["ATG", "CTG", "GTG", "TTG", "ATC", "ATA", "ATT", "ACG", "AAG", "AGG"]

stop_codons = ["TAG", "TAA", "TGA"]


# Keep track of the longest transcript for each noncoding gene, append this to transcript list later
longest_noncoding = {}


tran_count = 0
# This section is used to gather all cds regions, convert them to genomic regions and store them in a dictionary to check against later (all transcript contribute to this not just those
# in the transcript list)
genomic_cds_dict = {}
tree_dict = {}
for transcript in master_dict:
    # print (transcript, master_dict[transcript]["tran_type"])
    tran_count += 1
    if "seq" not in master_dict[transcript]:
        continue
    chrom = master_dict[transcript]["chrom"]
    if chrom not in genomic_cds_dict:
        genomic_cds_dict[chrom] = []
    if "cds_start" in master_dict[transcript]:
        cds_start = master_dict[transcript]["cds_start"]
        cds_stop = master_dict[transcript]["cds_stop"]
        if cds_start != "NULL":
            cds_pos = []
            for i in range(cds_start, cds_stop + 1):
                cds_pos.append(i)

            for tup in master_dict[transcript]["cds"]:
                if tup[0] != tup[1]:
                    if tup not in genomic_cds_dict[chrom]:
                        genomic_cds_dict[chrom].append(tup)

print("genomic cds dict built")
print(list(genomic_cds_dict))
for chrom in genomic_cds_dict:
    tree_dict[chrom] = IntervalTree.from_tuples(genomic_cds_dict[chrom])

# print (list(tree_dict))


connection = sqlite3.connect("{}.sqlite".format(organism))
cursor = connection.cursor()
cursor.execute(
    "CREATE TABLE IF NOT EXISTS transcripts (transcript VARCHAR(50), gene VARCHAR(50), length INT(6), cds_start INT(6), cds_stop INT(6), sequence VARCHAR(50000), strand CHAR(1), stop_list VARCHAR(10000), start_list VARCHAR(10000), exon_junctions VARCHAR(1000), tran_type INT(1), gene_type INT(1), principal INT(1), version INT(2),gc INT(3),five_gc INT(3), cds_gc INT(3), three_gc INT(3), chrom VARCHAR(20));"
)
cursor.execute(
    "CREATE TABLE IF NOT EXISTS coding_regions (transcript VARCHAR(50), coding_start INT(6), coding_stop INT(6));"
)
cursor.execute(
    "CREATE TABLE IF NOT EXISTS exons (transcript VARCHAR(50), exon_start INT(6), exon_stop INT(6));"
)
cursor.execute(
    "CREATE TABLE IF NOT EXISTS uorf (transcript VARCHAR(300), start_codon VARCHAR(10), length INT(6), start INT(6), stop INT(6), cds_coverage FLOAT(20));"
)
cursor.execute(
    "CREATE TABLE IF NOT EXISTS ouorf (transcript VARCHAR(300), start_codon VARCHAR(10), length INT(6), start INT(6), stop INT(6), cds_coverage FLOAT(20));"
)
cursor.execute(
    "CREATE TABLE IF NOT EXISTS cds (transcript VARCHAR(300), start_codon VARCHAR(10), length INT(6), start INT(6), stop INT(6), cds_coverage FLOAT(20));"
)
cursor.execute(
    "CREATE TABLE IF NOT EXISTS nested (transcript VARCHAR(300), start_codon VARCHAR(10), length INT(6), start INT(6), stop INT(6), cds_coverage FLOAT(20));"
)
cursor.execute(
    "CREATE TABLE IF NOT EXISTS odorf (transcript VARCHAR(300), start_codon VARCHAR(10), length INT(6), start INT(6), stop INT(6), cds_coverage FLOAT(20));"
)
cursor.execute(
    "CREATE TABLE IF NOT EXISTS dorf (transcript VARCHAR(300), start_codon VARCHAR(10), length INT(6), start INT(6), stop INT(6), cds_coverage FLOAT(20));"
)
cursor.execute(
    "CREATE TABLE IF NOT EXISTS minusone(transcript VARCHAR(300), start_codon VARCHAR(10), length INT(6), start INT(6), stop INT(6), cds_coverage FLOAT(20));"
)
cursor.execute(
    "CREATE TABLE IF NOT EXISTS readthrough (transcript VARCHAR(300), start_codon VARCHAR(10), length INT(6), start INT(6), stop INT(6), cds_coverage FLOAT(20));"
)
cursor.execute(
    "CREATE TABLE IF NOT EXISTS plusone (transcript VARCHAR(300), start_codon VARCHAR(10), length INT(6), start INT(6), stop INT(6), cds_coverage FLOAT(20));"
)
cursor.execute(
    "CREATE TABLE IF NOT EXISTS noncoding (transcript VARCHAR(300), start_codon VARCHAR(10), length INT(6), start INT(6), stop INT(6), cds_coverage FLOAT(20));"
)
cursor.execute(
    "CREATE TABLE IF NOT EXISTS extension (transcript VARCHAR(300), start_codon VARCHAR(10), length INT(6), start INT(6), stop INT(6), cds_coverage FLOAT(20));"
)
cursor.execute(
    "CREATE TABLE IF NOT EXISTS inframe_stop (transcript VARCHAR(300), start_codon VARCHAR(10), length INT(6), start INT(6), stop INT(6), cds_coverage FLOAT(20));"
)
connection.commit()


print("Finding ORFs")
transcript_count = 0
total_transcripts = len(list(master_dict))
for transcript in master_dict:
    # print ("transcript",transcript)
    # if transcript != "ENST00000316448":
    # 	continue
    transcript_count += 1
    if transcript_count % 100 == 0:
        print("Transcripts complete: {}/{}".format(transcript_count, total_transcripts))
    if "seq" not in master_dict[transcript]:
        print("transcript {} has no sequence".format(transcript))
        continue
    seq = master_dict[transcript]["seq"]
    cds_start = "NULL"
    cds_stop = "NULL"
    transcript_len = len(seq)
    if "cds_start" in master_dict[transcript]:
        cds_start = master_dict[transcript]["cds_start"]
        cds_stop = master_dict[transcript]["cds_stop"]

    # Find out what regions of this transcript overlap with any other coding regions
    coding_positions = []
    if cds_start != "NULL":
        # If this is a coding transcript don't bother checking the CDS
        for i in range(cds_start, cds_stop):
            coding_positions.append(i)
        # check 5' leader
        chrom, pos_list = tran_to_genome(transcript, 0, cds_start, master_dict)
        for i in range(0, cds_start):
            genomic_pos = pos_list[i]
            overlap = tree_dict[chrom][genomic_pos]
            if len(overlap) != 0:
                coding_positions.append(i)
        # check 3' trailer
        chrom, pos_list = tran_to_genome(
            transcript, cds_stop, transcript_len, master_dict
        )
        for i in range(cds_stop, transcript_len + 1):
            # print ("i",i)
            genomic_pos = pos_list[i - cds_stop]
            # print ("genomic position",genomic_pos)
            overlap = tree_dict[chrom][genomic_pos]
            if len(overlap) != 0:
                # print ("overlap not empty appending i",overlap)
                coding_positions.append(i)
    else:
        # check entire transcript
        chrom, pos_list = tran_to_genome(transcript, 0, transcript_len, master_dict)
        for i in range(0, transcript_len):
            genomic_pos = pos_list[i]
            overlap = tree_dict[chrom][genomic_pos]
            if len(overlap) != 0:
                coding_positions.append(i)
    coding_positions_tuple = to_ranges(coding_positions)
    coding_dict[transcript] = coding_positions_tuple
    coding_positions = set(coding_positions)
    # if this is a coding transcript find the minusone, readhtrough, plusone co-ordinates
    if cds_start != "NULL":
        # print ("transcript", transcript)
        # if pseudo_utr_len != 0:
        # 	cds_stop -= 3 # take 3 from stop so we can match it with orf_stop, do it here rather than above in case cds_stop is null
        recoding_dict = {2: "minusone", 0: "readthrough", 1: "plusone"}
        for addition in recoding_dict:
            orftype = recoding_dict[addition]
            for i in range(cds_stop + addition, transcript_len, 3):
                if seq[i : i + 3] in stop_codons:
                    # orf_seq = seq[cds_stop:i+3]
                    orf_start = cds_stop
                    if orftype == "readthrough":
                        orf_start -= 2
                    if orftype == "plusone":
                        orf_start -= 1
                    orf_stop = i + 3  # +2 so it refers to the end of the stop codon
                    start_codon = None
                    length = (i + 3) - cds_stop
                    orf_pos_list = []
                    # determine how many nucleotides in this orf overlap with an annotated coding region
                    cds_cov_count = 0.0
                    for position in range(orf_start, orf_stop):
                        orf_pos_list.append(position)
                    for pos in range(orf_start, orf_stop + 1):
                        if pos in coding_positions:
                            cds_cov_count += 1
                    cds_cov = cds_cov_count / length
                    # print ("orftype, start, stop", orftype, orf_start, orf_stop)
                    cursor.execute(
                        "INSERT INTO {} VALUES('{}','{}',{},{},{},{});".format(
                            orftype,
                            transcript,
                            start_codon,
                            length,
                            orf_start,
                            orf_stop,
                            cds_cov,
                        )
                    )
                    break
        # sys.exit()
    for frame in [0, 1, 2]:
        for i in range(frame, transcript_len, 3):
            if seq[i : i + 3] in start_codons:
                for x in range(i, transcript_len, 3):
                    if seq[x : x + 3] in stop_codons:
                        # orf_seq = seq[i:x+3]
                        orf_start = i + 1
                        orf_stop = x + 3  # +2 so it refers to the end of the stop codon
                        start_codon = seq[i : i + 3]
                        length = (x + 3) - i
                        orf_pos_list = []
                        # determine how many nucleotides in this orf overlap with an annotated coding region
                        cds_cov_count = 0.0
                        for pos in range(orf_start, orf_stop + 1):
                            if pos in coding_positions:
                                cds_cov_count += 1
                        cds_cov = float(cds_cov_count) / float(length)
                        # Now determine orf type
                        if cds_start == "NULL":
                            orftype = "noncoding"
                        else:
                            # print ("cds start is not null :{}:{}".format(cds_start,cds_stop))
                            # print ("orf start, orf stop", orf_start, orf_stop)
                            if orf_start == cds_start and orf_stop == cds_stop:
                                orftype = "cds"
                                # print ("orf type is cds")
                            elif orf_start < cds_start and orf_stop == cds_stop:
                                orftype = "extension"
                                # special case for extensions, we only take from the orf_start to the cds_start, and re-calculate cds coverage
                                orf_stop = cds_start
                                cds_cov_count = 0.0
                                for pos in range(orf_start, orf_stop + 1):
                                    if pos in coding_positions:
                                        cds_cov_count += 1
                                cds_cov = float(cds_cov_count) / float(length)
                                # orf_seq = seq[orf_start:cds_start]
                                length = cds_start - orf_start
                                # print ("orf type is extension")
                            elif orf_start < cds_start and orf_stop <= cds_start:
                                orftype = "uorf"
                                # print ("orf type is uorf")
                            elif orf_start < cds_start and orf_stop > cds_start:
                                orftype = "ouorf"
                                # print ("orf type is ouorf")
                                # sys.exit()
                            elif (
                                orf_start >= cds_start
                                and orf_start <= cds_stop
                                and orf_stop <= cds_stop
                            ):
                                if orf_stop == cds_stop:
                                    break
                                # print ("Tran, cds_start, cds_stop, orf_start, orf_stop, tranlen",tran, cds_start, cds_stop, orf_start, orf_stop,transcript_len)
                                if (
                                    orf_stop < transcript_len
                                    and orf_stop % 3 == cds_stop % 3
                                ) or (
                                    cds_start != 1
                                    and orf_stop % 3 == (cds_start + 2) % 3
                                ):
                                    # print ("Transcript {} has an inframe stop codon".format(transcript))
                                    break
                                orftype = "nested"
                                # print ("orf type is nested")
                            elif (
                                orf_start >= cds_start
                                and orf_start <= cds_stop
                                and orf_stop > cds_stop
                            ):
                                orftype = "odorf"
                                # print ("orftype is odorf")
                            elif orf_start > cds_stop and orf_stop > cds_stop:
                                orftype = "dorf"
                                # print ("orftype is dorf")
                            # if orf_stop > cds_start and orf_stop < cds_stop:
                            # 	if (orf_stop+1)%3 == cds_start%3:
                            # 		orftype = "inframe_stop"
                            # 		print ("inframe stop, transcript, orf_stop", transcript, orf_stop)
                            # 		sys.exit()
                            # 		if transcript not in orf_dict:
                            # 			orf_dict[orftype][transcript] = []
                            # 	#print ("some weird stop or something")
                        cursor.execute(
                            "INSERT INTO {} VALUES('{}','{}',{},{},{},{});".format(
                                orftype,
                                transcript,
                                start_codon,
                                length,
                                orf_start,
                                orf_stop,
                                cds_cov,
                            )
                        )
                        break
# Used to keep track of the codons at cds_start and cds_stop positions,
# If there is an issue with the cds co-ordinates the starts and stops counts will
# be much lower than the other count, start with 1 to prevent division by 0
nuc_dict = {"stops": {"stops": 1, "other": 0}, "starts": {"starts": 1, "other": 0}}


def calcgc(seq):
    if seq == "":
        return "NULL"
    g_count = 0
    c_count = 0
    a_count = 0
    t_count = 0
    for char in seq:
        if char == "A":
            a_count += 1
        if char == "T":
            t_count += 1
        if char == "G":
            g_count += 1
        if char == "C":
            c_count += 1
        gc = ((g_count + c_count) / float(len(seq))) * 100
    return round(gc, 2)


for transcript in master_dict:
    # print ("transcripts", transcript)
    length = master_dict[transcript]["length"]
    cds_start = master_dict[transcript]["cds_start"]
    cds_stop = master_dict[transcript]["cds_stop"]
    seq = master_dict[transcript]["seq"]
    strand = master_dict[transcript]["strand"]
    chrom = master_dict[transcript]["chrom"]
    gene = master_dict[transcript]["gene_name"]
    gc = calcgc(seq)
    five_gc = "NULL"
    cds_gc = "NULL"
    three_gc = "NULL"
    if cds_start != "NULL":
        five_gc = calcgc(seq[:cds_start])
        cds_gc = calcgc(seq[cds_start:cds_stop])
        three_gc = calcgc(seq[cds_stop:])
        # check that the nucleotide cds_start points to is the first of the start codon
        # take one becase cds_start is 1 based but python indexing is 0 based
        start_nuc = seq[cds_start - 1 : cds_start + 2]
        # print ("start nuc",start_nuc)
        if start_nuc == "ATG":
            nuc_dict["starts"]["starts"] += 1
        else:
            nuc_dict["starts"]["other"] += 1
        stop_nuc = seq[cds_stop - 3 : cds_stop]
        # print ("stop_nuc",stop_nuc)
        if stop_nuc in ["TAG", "TAA", "TGA"]:
            nuc_dict["stops"]["stops"] += 1
        else:
            nuc_dict["stops"]["other"] += 1
    tran_type = master_dict[transcript]["tran_type"]
    if coding_genes_dict[gene] == True:
        gene_type = 1
    else:
        gene_type = 0
    # print ("tran type before",tran_type)
    if tran_type == "coding":
        tran_type = 1
    else:
        tran_type = 0
    # print ("tran type after",tran_type)
    start_list = str(master_dict[transcript]["start_list"]).replace(" ", "").strip("[]")
    stop_list = str(master_dict[transcript]["stop_list"]).replace(" ", "").strip("[]")
    exon_junctions = (
        str(master_dict[transcript]["exon_junctions"]).replace(" ", "").strip("[]")
    )
    principal = master_dict[transcript]["principal"]
    version = master_dict[transcript]["version"]
    # print (master_dict[transcript])
    # print (tran_type)
    # print (gene_type)
    # print (principal)
    # print (version)
    # print ("INSERT INTO transcripts VALUES('{}','{}',{},{},{},'{}','{}','{}','{}','{}',{},{},{},{});".format(transcript, gene, length, cds_start, cds_stop, seq, strand,stop_list, start_list, exon_junctions, tran_type,gene_type,principal,version))
    cursor.execute(
        "INSERT INTO transcripts VALUES('{}','{}',{},{},{},'{}','{}','{}','{}','{}',{},{},{},{},{},{},{},{},'{}');".format(
            transcript,
            gene,
            length,
            cds_start,
            cds_stop,
            seq,
            strand,
            stop_list,
            start_list,
            exon_junctions,
            tran_type,
            gene_type,
            principal,
            version,
            gc,
            five_gc,
            cds_gc,
            three_gc,
            chrom,
        )
    )

    for tup in master_dict[transcript]["exon"]:
        cursor.execute(
            "INSERT INTO exons VALUES('{}',{},{});".format(transcript, tup[0], tup[1])
        )
    if transcript in coding_dict:
        for tup in coding_dict[transcript]:
            cursor.execute(
                "INSERT INTO coding_regions VALUES('{}',{},{});".format(
                    transcript, tup[0], tup[1]
                )
            )

connection.commit()
connection.close()

print("delim", delimiters)
if (nuc_dict["starts"]["other"] / nuc_dict["starts"]["starts"]) > 0.05:
    print(
        "Warning: {} transcripts do not have a an AUG at the CDS start position".format(
            nuc_dict["starts"]["other"]
        )
    )
if (nuc_dict["stops"]["other"] / nuc_dict["stops"]["stops"]) > 0.05:
    print(
        "Warning: {} transcripts do not have a an stop codon at the CDS stop position".format(
            nuc_dict["stops"]["other"]
        )
    )
if len(notinannotation) > 0:
    print(
        "Warning: {} transcripts were in the fasta file, but not the annotation file, these will be discarded".format(
            len(notinannotation)
        )
    )