{
"cells": [
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# Conformation Layer\n",
"\n",
"@Author: 吴炜坤\n",
"\n",
"@email：weikun.wu@xtalpi.com weikunwu@163.com\n"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"在前一个章节中，我们已经介绍了在Rosetta中，Residue对象是描述蛋白质的基本单元，许多独立的Residue被用于描述蛋白质的几何结构和高级构象。\n",
"这些细微的构象变化的度量就是由**原子间的键长、键角，二面角等一系列的具体参数构成。**在PyRosetta中, Pose中这些几何构象的参数由Conformation对象负责记录。"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### 一、蛋白质的几何构象\n",
"蛋白质是由多个氨基酸通过脱水缩合的方式形成肽键共价连接。因此对于天然氨基酸而言。骨架最重要的两个二面角就是phi和psi角，而omega由于肽键平面一般处于0或180°附近。对于不同的氨基酸侧链，每个Residue含有若干个chi角。这些二面角组成了Rosetta对构象采样的基本几何参数。"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"<center><img src=\"./img/phipsiomega.png\" width = \"600\" height = \"200\" align=center /></center>\n",
"(图片来源: https://github.com/RosettaCommons/PyRosetta.notebooks/tree/master/student-notebooks)"
]
},
{
"cell_type": "code",
"execution_count": 1,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"PyRosetta-4 2021 [Rosetta PyRosetta4.conda.mac.cxx11thread.serialization.python37.Release 2021.31+release.c7009b3115c22daa9efe2805d9d1ebba08426a54 2021-08-07T10:04:12] retrieved from: http://www.pyrosetta.org\n",
"(C) Copyright Rosetta Commons Member Institutions. Created in JHU by Sergey Lyskov and PyRosetta Team.\n",
"\u001b[0mcore.init: {0} \u001b[0mChecking for fconfig files in pwd and ./rosetta/flags\n",
"\u001b[0mcore.init: {0} \u001b[0mRosetta version: PyRosetta4.conda.mac.cxx11thread.serialization.python37.Release r292 2021.31+release.c7009b3115c c7009b3115c22daa9efe2805d9d1ebba08426a54 http://www.pyrosetta.org 2021-08-07T10:04:12\n",
"\u001b[0mcore.init: {0} \u001b[0mcommand: PyRosetta -ex1 -ex2aro -database /opt/miniconda3/lib/python3.7/site-packages/pyrosetta/database\n",
"\u001b[0mbasic.random.init_random_generator: {0} \u001b[0m'RNG device' seed mode, using '/dev/urandom', seed=265248943 seed_offset=0 real_seed=265248943 thread_index=0\n",
"\u001b[0mbasic.random.init_random_generator: {0} \u001b[0mRandomGenerator:init: Normal mode, seed=265248943 RG_type=mt19937\n",
"\u001b[0mcore.chemical.GlobalResidueTypeSet: {0} \u001b[0mFinished initializing fa_standard residue type set. Created 983 residue types\n",
"\u001b[0mcore.chemical.GlobalResidueTypeSet: {0} \u001b[0mTotal time to initialize 0.67614 seconds.\n",
"\u001b[0mcore.import_pose.import_pose: {0} \u001b[0mFile './data/4jfx_peptide.pdb' automatically determined to be of type PDB\n"
]
}
],
"source": [
"# 读取多肽的PDB结构\n",
"from pyrosetta import init, pose_from_pdb\n",
"init()\n",
"pose = pose_from_pdb('./data/4jfx_peptide.pdb')"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"#### 1.1 基本二面角几何参数\n",
"从pyrosetta中获取这些基本几何参数的方式非常简单:"
]
},
{
"cell_type": "code",
"execution_count": 2,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"-148.8025511852094 157.0624491251048 157.0624491251048\n"
]
}
],
"source": [
"# 获取第3号氨基酸的骨架二面角:\n",
"phi = pose.phi(3)\n",
"psi = pose.psi(3)\n",
"omega = pose.psi(3)\n",
"print(phi, psi, omega)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"也可以直接获取侧链的二面角参数: \n",
"\n",
"pose.chi($\\chi$角编号, Residue的pose编号)即可获取。"
]
},
{
"cell_type": "code",
"execution_count": 3,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"3\n"
]
}
],
"source": [
"# 获取第三号残基的chi角信息: pose.chi(chi_id:int, residue_id:int)\n",
"chi_num = len(pose.residue(3).chi_atoms())\n",
"print(chi_num)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"残基中共有3个$\\chi$二面角。"
]
},
{
"cell_type": "code",
"execution_count": 4,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"70.4233120899352\n",
"81.97278776078356\n",
"-2.023205011879734e-14\n"
]
}
],
"source": [
"# 打印每个chi二面角的\n",
"for chi_id in range(1, chi_num+1):\n",
" chi_angle = pose.chi(chi_id, 3)\n",
" print(chi_angle)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"#### 1.2 调整二面角几何参数\n",
"Pose对象中内置的几个函数非常方便地可以用于调整几何构象: set_phi, set_psi, set_omega, set_chi。"
]
},
{
"cell_type": "code",
"execution_count": 5,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"True"
]
},
"execution_count": 5,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"# 调整骨架二面角\n",
"pose.set_phi(seqpos=3, setting=-150)\n",
"pose.set_phi(seqpos=3, setting=170)\n",
"pose.dump_pdb('./data/4jfx_peptide_conf0.pdb')"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"调整构象后的多肽构象直观感受:\n",
"<center><img src=\"./img/conf.png\" width = \"500\" height = \"200\" align=center /></center>\n",
"(图片来源: 晶泰科技团队)"
]
},
{
"cell_type": "code",
"execution_count": 6,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"True"
]
},
"execution_count": 6,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"# 调整3号氨基酸的侧链chi1角的角度;\n",
"pose.set_chi(chino=1, seqpos=3, setting=60)\n",
"pose.dump_pdb('./data/4jfx_peptide_chi_conf0.pdb')\n",
"pose.set_chi(chino=1, seqpos=3, setting=-60)\n",
"pose.dump_pdb('./data/4jfx_peptide_chi_conf1.pdb')\n",
"pose.set_chi(chino=1, seqpos=3, setting=180)\n",
"pose.dump_pdb('./data/4jfx_peptide_chi_conf2.pdb')"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"<center><img src=\"./img/chi_conf.png\" width = \"500\" height = \"200\" align=center /></center>\n",
"(图片来源: 晶泰科技团队)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### 二、化学键几何参数\n",
"除了对构象变化依赖影响最大的二面角参数，局部的化学键键长和键角信息也储存在Conformation对象中。\n",
"为了定位原子的信息，首先需要构建atom identifier对象，相当于创建一个ID卡，让Rosetta知道我们指定的原子是位于哪个氨基酸中的。通过AtomID，提供残基号，原子号，就可以创建atom identifier对象"
]
},
{
"cell_type": "code",
"execution_count": 7,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
" atomno= 1 rsd= 3 \n",
" atomno= 2 rsd= 3 \n",
" atomno= 3 rsd= 3 \n",
" atomno= 4 rsd= 3 \n"
]
}
],
"source": [
"# 获取原子间的键长、键角信息前需要构建atom identifier objects\n",
"from pyrosetta.rosetta.core.id import AtomID\n",
"atom1 = AtomID(atomno_in=1, rsd_in=3) # 3号残基的第一个原子\n",
"atom2 = AtomID(atomno_in=2, rsd_in=3) # 3号残基的第二个原子\n",
"atom3 = AtomID(atomno_in=3, rsd_in=3) # 3号残基的第三个原子\n",
"atom4 = AtomID(atomno_in=4, rsd_in=3) # 3号残基的第四个原子\n",
"print(atom1)\n",
"print(atom2)\n",
"print(atom3)\n",
"print(atom4)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"知道原子的ID后，就可以通过conformation对象来获取键长、键角等数据了。但一般这些参数在Rosetta中键长和键角都设定为理想值，可以极大减少蛋白质构象的采样自由度空间。但注意的是，获取的键长键角必须是有“物理连接的”。"
]
},
{
"cell_type": "code",
"execution_count": 8,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"键长:1.4632750937537378, 键角:1.9840915800459624\n"
]
}
],
"source": [
"# 通过conformation层获取键长数据\n",
"bond_length = pose.conformation().bond_length(atom1, atom2)\n",
"\n",
"# 通过conformation层获取键角数据(弧度)\n",
"bond_angle = pose.conformation().bond_angle(atom1, atom2, atom3)\n",
"\n",
"print(f'键长:{bond_length}, 键角:{bond_angle}')"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"同样原子间的键长和键角也是可以被调整的:"
]
},
{
"cell_type": "code",
"execution_count": 9,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"1.5 3.4\n"
]
}
],
"source": [
"# 设置新的值:\n",
"pose.conformation().set_bond_length(atom1, atom2, setting=1.5) # 设置键长\n",
"pose.conformation().set_bond_angle(atom1, atom2, atom3, setting=3.4) # 设置键角，弧度，而非角度\n",
"\n",
"# 查看新的值设定情况:\n",
"new_bond_length = pose.conformation().bond_length(atom1, atom2)\n",
"new_bond_angle = pose.conformation().bond_angle(atom1, atom2, atom3)\n",
"print(new_bond_length, new_bond_angle)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### 三、原子xyz坐标\n",
"原子坐标的修改需要获取residue对象，并获取原子ID(atom identifier objects)。通过pose.set_xyz函数设定新的xyz坐标, 但用户一般不需要”显式“地修改原子坐标, 除非你明白这样操作的意义。此处我们沿着3个坐标轴平移所有原子3个埃的距离。"
]
},
{
"cell_type": "code",
"execution_count": 10,
"metadata": {},
"outputs": [],
"source": [
"# 原子坐标的修改（一般不需要这样操作）\n",
"from pyrosetta.rosetta.numeric import xyzVector_double_t\n",
"\n",
"# 对所有氨基酸的所有原子的x坐标乘上一个负号:\n",
"for residue_id in range(1, pose.total_residue()+1):\n",
" residue = pose.residue(residue_id) # 获取residue对象\n",
" for atom_id, atom in enumerate(residue.atoms()):\n",
" x, y, z = atom.xyz()\n",
" \n",
" # 镜像处理xyz坐标:\n",
" trans_xyz = xyzVector_double_t(x+3, y+3, z+3) # 平移+3埃.\n",
" atom_index = AtomID(atom_id+1, residue_id) # 3号氨基酸的第x个原子的id\n",
" pose.set_xyz(atom_index, trans_xyz) # 设置xyz坐标"
]
},
{
"cell_type": "code",
"execution_count": 11,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"True"
]
},
"execution_count": 11,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"pose.dump_pdb('./data/trans.pdb')"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"<center><img src=\"./img/trans_xyz.png\" width = \"600\" height = \"200\" align=center /></center>\n",
"(图片来源: 晶泰科技团队)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### 四、 蛋白二级结构信息\n",
"Rosetta中的二级结构信息来源于DSSP的计算。通过get_secstruct函数即可获取。"
]
},
{
"cell_type": "code",
"execution_count": 12,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"\u001b[0mprotocols.DsspMover: {0} \u001b[0mLLLLLLLL\n",
"LLLLLLLL\n"
]
}
],
"source": [
"# 通过DSSP获取二级结构信息\n",
"from pyrosetta.rosetta.protocols.membrane import get_secstruct\n",
"ss = ''.join(get_secstruct(pose))\n",
"print(ss)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": []
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### 五、理想化初始拓扑数据\n",
"在晶体中，可能会存在一些非理想二面角、键长、键角等。可以通过IdealizeMover进行修复。"
]
},
{
"cell_type": "code",
"execution_count": 13,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"\u001b[0mprotocols.idealize.IdealizeMover: {0} \u001b[0mtotal atompairs: 0\n",
"\u001b[0mprotocols.idealize: {0} \u001b[0mlastjumpmin: 1\n",
"\u001b[0mprotocols.idealize: {0} \u001b[0mpremin: (pos,rmsd,avg-bb,max-bb,avg-chi,max-chi,score) 2 N 0.075 0.000 0.000 0.000 0.000 1.416\n",
"\u001b[0mprotocols.idealize: {0} \u001b[0mpostmin: (pos,rmsd,avg-bb,max-bb,avg-chi,max-chi,score) 2 N 0.039 4.106 5.979 2.375 4.050 0.283\n",
"\u001b[0mprotocols.idealize: {0} \u001b[0m\n",
"------------------------------------------------------------\n",
" Scores Weight Raw Score Wghtd.Score\n",
"------------------------------------------------------------\n",
" pro_close 0.500 0.000 0.000\n",
" dslf_ss_dst 0.500 0.000 0.000\n",
" dslf_cs_ang 2.000 0.000 0.000\n",
" coordinate_constraint 0.010 28.336 0.283\n",
"---------------------------------------------------\n",
" Total weighted score: 0.283\n",
"\u001b[0mprotocols.idealize: {0} \u001b[0mpremin: (pos,rmsd,avg-bb,max-bb,avg-chi,max-chi,score) 3 Y 2.619 2.531 5.979 0.950 4.050 1095.920\n",
"\u001b[0mprotocols.idealize: {0} \u001b[0mpostmin: (pos,rmsd,avg-bb,max-bb,avg-chi,max-chi,score) 3 Y 0.656 46.121 77.515 28.485 109.577 56.651\n",
"\u001b[0mprotocols.idealize: {0} \u001b[0m\n",
"------------------------------------------------------------\n",
" Scores Weight Raw Score Wghtd.Score\n",
"------------------------------------------------------------\n",
" pro_close 0.500 0.000 0.000\n",
" dslf_ss_dst 0.500 0.000 0.000\n",
" dslf_cs_ang 2.000 0.000 0.000\n",
" coordinate_constraint 0.010 5665.102 56.651\n",
"---------------------------------------------------\n",
" Total weighted score: 56.651\n",
"\u001b[0mprotocols.idealize: {0} \u001b[0mpremin: (pos,rmsd,avg-bb,max-bb,avg-chi,max-chi,score) 4 V 0.657 44.944 118.620 23.737 109.577 56.780\n",
"\u001b[0mprotocols.idealize: {0} \u001b[0mpostmin: (pos,rmsd,avg-bb,max-bb,avg-chi,max-chi,score) 4 V 0.657 44.944 118.629 23.737 109.577 56.747\n",
"\u001b[0mprotocols.idealize: {0} \u001b[0m\n",
"------------------------------------------------------------\n",
" Scores Weight Raw Score Wghtd.Score\n",
"------------------------------------------------------------\n",
" pro_close 0.500 0.000 0.000\n",
" dslf_ss_dst 0.500 0.000 0.000\n",
" dslf_cs_ang 2.000 0.000 0.000\n",
" coordinate_constraint 0.010 5674.731 56.747\n",
"---------------------------------------------------\n",
" Total weighted score: 56.747\n",
"\u001b[0mprotocols.idealize: {0} \u001b[0mpremin: (pos,rmsd,avg-bb,max-bb,avg-chi,max-chi,score) 5 V 0.654 34.395 118.629 20.346 109.577 57.267\n",
"\u001b[0mprotocols.idealize: {0} \u001b[0mpostmin: (pos,rmsd,avg-bb,max-bb,avg-chi,max-chi,score) 5 V 0.656 34.377 118.361 20.353 109.577 56.568\n",
"\u001b[0mprotocols.idealize: {0} \u001b[0m\n",
"------------------------------------------------------------\n",
" Scores Weight Raw Score Wghtd.Score\n",
"------------------------------------------------------------\n",
" pro_close 0.500 0.000 0.000\n",
" dslf_ss_dst 0.500 0.000 0.000\n",
" dslf_cs_ang 2.000 0.000 0.000\n",
" coordinate_constraint 0.010 5656.769 56.568\n",
"---------------------------------------------------\n",
" Total weighted score: 56.568\n",
"\u001b[0mprotocols.idealize: {0} \u001b[0mpremin: (pos,rmsd,avg-bb,max-bb,avg-chi,max-chi,score) 6 T 0.667 27.842 118.361 15.830 109.577 60.651\n",
"\u001b[0mprotocols.idealize: {0} \u001b[0mpostmin: (pos,rmsd,avg-bb,max-bb,avg-chi,max-chi,score) 6 T 0.655 27.895 118.453 15.832 109.577 56.600\n",
"\u001b[0mprotocols.idealize: {0} \u001b[0m\n",
"------------------------------------------------------------\n",
" Scores Weight Raw Score Wghtd.Score\n",
"------------------------------------------------------------\n",
" pro_close 0.500 0.000 0.000\n",
" dslf_ss_dst 0.500 0.000 0.000\n",
" dslf_cs_ang 2.000 0.000 0.000\n",
" coordinate_constraint 0.010 5660.035 56.600\n",
"---------------------------------------------------\n",
" Total weighted score: 56.600\n",
"\u001b[0mprotocols.idealize: {0} \u001b[0mpremin: (pos,rmsd,avg-bb,max-bb,avg-chi,max-chi,score) 8 A 0.670 26.155 118.453 15.832 109.577 57.959\n",
"\u001b[0mprotocols.idealize: {0} \u001b[0mpostmin: (pos,rmsd,avg-bb,max-bb,avg-chi,max-chi,score) 8 A 0.670 26.155 118.453 15.832 109.577 57.957\n",
"\u001b[0mprotocols.idealize: {0} \u001b[0m\n",
"------------------------------------------------------------\n",
" Scores Weight Raw Score Wghtd.Score\n",
"------------------------------------------------------------\n",
" pro_close 0.500 0.000 0.000\n",
" dslf_ss_dst 0.500 0.000 0.000\n",
" dslf_cs_ang 2.000 0.000 0.000\n",
" coordinate_constraint 0.010 5795.692 57.957\n",
"---------------------------------------------------\n",
" Total weighted score: 57.957\n",
"\u001b[0mprotocols.idealize: {0} \u001b[0mpremin: (pos,rmsd,avg-bb,max-bb,avg-chi,max-chi,score) 7 Y 0.796 22.061 118.453 11.874 109.577 71.358\n",
"\u001b[0mprotocols.idealize: {0} \u001b[0mpostmin: (pos,rmsd,avg-bb,max-bb,avg-chi,max-chi,score) 7 Y 0.796 22.060 118.455 11.875 109.577 71.337\n",
"\u001b[0mprotocols.idealize: {0} \u001b[0m\n",
"------------------------------------------------------------\n",
" Scores Weight Raw Score Wghtd.Score\n",
"------------------------------------------------------------\n",
" pro_close 0.500 0.000 0.000\n",
" dslf_ss_dst 0.500 0.000 0.000\n",
" dslf_cs_ang 2.000 0.000 0.000\n",
" coordinate_constraint 0.010 7133.705 71.337\n",
"---------------------------------------------------\n",
" Total weighted score: 71.337\n",
"\u001b[0mprotocols.idealize: {0} \u001b[0mpremin: (pos,rmsd,avg-bb,max-bb,avg-chi,max-chi,score) 1 G 0.796 21.383 118.455 11.875 109.577 71.315\n",
"\u001b[0mprotocols.idealize: {0} \u001b[0mpostmin: (pos,rmsd,avg-bb,max-bb,avg-chi,max-chi,score) 1 G 0.796 21.383 118.457 11.875 109.577 71.307\n",
"\u001b[0mprotocols.idealize: {0} \u001b[0m\n",
"------------------------------------------------------------\n",
" Scores Weight Raw Score Wghtd.Score\n",
"------------------------------------------------------------\n",
" pro_close 0.500 0.000 0.000\n",
" dslf_ss_dst 0.500 0.000 0.000\n",
" dslf_cs_ang 2.000 0.000 0.000\n",
" coordinate_constraint 0.010 7130.685 71.307\n",
"---------------------------------------------------\n",
" Total weighted score: 71.307\n",
"\u001b[0mprotocols.idealize: {0} \u001b[0mpre-finalmin: (pos,rmsd,avg-bb,max-bb,avg-chi,max-chi,score) 0.796 21.383 118.457 11.875 109.577 71.307\n",
"\u001b[0mprotocols.idealize: {0} \u001b[0mpost-finalmin: (pos,rmsd,avg-bb,max-bb,avg-chi,max-chi,score) 0.796 21.383 118.459 11.875 109.576 71.292\n",
"\u001b[0mprotocols.idealize: {0} \u001b[0m\n",
"------------------------------------------------------------\n",
" Scores Weight Raw Score Wghtd.Score\n",
"------------------------------------------------------------\n",
" pro_close 0.500 0.000 0.000\n",
" dslf_ss_dst 0.500 0.000 0.000\n",
" dslf_cs_ang 2.000 0.000 0.000\n",
" coordinate_constraint 0.010 7129.185 71.292\n",
"---------------------------------------------------\n",
" Total weighted score: 71.292\n",
"\u001b[0mprotocols.idealize: {0} \u001b[0m\n",
"\u001b[0mprotocols.idealize.IdealizeMover: {0} \u001b[0mRMS between original pose and idealised pose: 0.524355 CA RMSD, 0.811604 All-Atom RMSD,\n"
]
}
],
"source": [
"from pyrosetta.rosetta.protocols.idealize import IdealizeMover\n",
"# idealized\n",
"idm = IdealizeMover()\n",
"idm.apply(pose)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"理想化之后，的All-Atom RMSD发生了轻微的变化。"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### 习题与思考\n",
"1. 尝试下载PDB数据库中的一些数据，统计分析它们的omega角，看看其分布的范围区间？\n",
"2. 尝试统计每种氨基酸的Phi/Psi角的分布，它们是否有些什么差异？\n",
"3. 不同的二面角它们产生分布差异的原因是什么？"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": []
}
],
"metadata": {
"kernelspec": {
"display_name": "Python 3",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.7.10"
}
},
"nbformat": 4,
"nbformat_minor": 4
}