{"id":85614,"date":"2026-07-27T19:08:22","date_gmt":"2026-07-27T11:08:22","guid":{"rendered":"https:\/\/www.wsisp.com\/helps\/85614.html"},"modified":"2026-07-27T19:08:22","modified_gmt":"2026-07-27T11:08:22","slug":"alphafold%e5%8e%9f%e7%90%86%e4%b8%8e%e8%9b%8b%e7%99%bd%e8%b4%a8%e7%bb%93%e6%9e%84%e9%a2%84%e6%b5%8b%e5%ae%9e%e6%88%98","status":"publish","type":"post","link":"https:\/\/www.wsisp.com\/helps\/85614.html","title":{"rendered":"AlphaFold\u539f\u7406\u4e0e\u86cb\u767d\u8d28\u7ed3\u6784\u9884\u6d4b\u5b9e\u6218"},"content":{"rendered":"

AlphaFold\u539f\u7406\u4e0e\u86cb\u767d\u8d28\u7ed3\u6784\u9884\u6d4b\u5b9e\u6218<\/h2>\n

\u4e00\u3001\u5f15\u8a00<\/h3>\n

"\u86cb\u767d\u8d28\u6298\u53e0\u95ee\u9898"\u88ab\u79f0\u4e3a\u751f\u7269\u5b6650\u5e74\u672a\u89e3\u96be\u9898\u30022020\u5e74 DeepMind \u7684 AlphaFold2 \u5728 CASP14 \u4e2d\u4ee5\u539f\u5b50\u7ea7\u7cbe\u5ea6（\u4e2d\u4f4d\u8bef\u5dee0.96\u00c5）\u9884\u6d4b\u86cb\u767d\u8d283D\u7ed3\u6784，\u9707\u60ca\u79d1\u5b66\u754c\u30022024\u5e74 AlphaFold3 \u8fdb\u4e00\u6b65\u6269\u5c55\u5230\u6240\u6709\u751f\u7269\u5206\u5b50\u3002<\/p>\n

\u672c\u6587\u5c06\u6df1\u5165\u89e3\u6790 AlphaFold2 \u7684\u6838\u5fc3\u67b6\u6784：MSA\u5904\u7406\u3001Pair\u8868\u793a\u3001\u7ed3\u6784\u6a21\u5757（IPA）\u3001\u56de\u6536\u673a\u5236，\u5e76\u63d0\u4f9b\u5b8c\u6574\u63a8\u7406\u4ee3\u7801\u3002<\/p>\n

\u4e8c\u3001\u86cb\u767d\u8d28\u7ed3\u6784\u57fa\u7840<\/h3>\n

# \u86cb\u767d\u8d28\u753120\u79cd\u6c28\u57fa\u9178\u7ec4\u6210<\/span>
\nAMINO_ACIDS =<\/span> "ACDEFGHIKLMNPQRSTVWY"<\/span> # 20\u79cd\u6807\u51c6\u6c28\u57fa\u9178<\/span><\/p>\n

# \u7ed3\u6784\u5c42\u6b21:<\/span>
\n# \u5e8f\u5217(1D) \u2192 \u4e8c\u7ea7\u7ed3\u6784(\u03b1\u87ba\u65cb\/\u03b2\u6298\u53e0) \u2192 \u4e09\u7ea7\u7ed3\u6784(3D\u5750\u6807)<\/span>
\n# \u6bcf\u4e2a\u6b8b\u57fa\u7531\u4e3b\u94fe\u539f\u5b50(N, C\u03b1, C) + \u4fa7\u94fe\u539f\u5b50\u7ec4\u6210<\/span><\/p>\n

# \u76ee\u6807: \u7ed9\u5b9a\u6c28\u57fa\u9178\u5e8f\u5217 \u2192 \u9884\u6d4b\u6bcf\u4e2a\u539f\u5b50\u76843D\u5750\u6807<\/span><\/p>\n

\u4e09\u3001AlphaFold2 \u67b6\u6784<\/h3>\n

\u8f93\u5165\u5e8f\u5217: "MKFLILFNILV…"
\n \u2502
\n \u251c\u2500\u2500\u2192 [\u57fa\u56e0\u6570\u636e\u5e93\u641c\u7d22] \u2192 MSA (\u591a\u5e8f\u5217\u6bd4\u5bf9)
\n \u2502 \u2514\u2500\u2500 JackHMMER\/Mmseqs2 \u2192 \u540c\u6e90\u5e8f\u5217
\n \u2502
\n \u251c\u2500\u2500\u2192 [\u7ed3\u6784\u6570\u636e\u5e93\u641c\u7d22] \u2192 Templates (\u7ed3\u6784\u6a21\u677f)
\n \u2502 \u2514\u2500\u2500 HHSearch \u2192 \u5df2\u77e5\u7ed3\u6784
\n \u2502
\n \u25bc
\n\u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510
\n\u2502 Evoformer (48\u5c42) \u2502
\n\u2502 \u251c\u2500\u2500 MSA Stack \u2502 \u884c\u6ce8\u610f\u529b\/\u5217\u6ce8\u610f\u529b
\n\u2502 \u2502 \u251c\u2500\u2500 Row-wise Gated \u2502
\n\u2502 \u2502 \u251c\u2500\u2500 Column-wise Gated \u2502
\n\u2502 \u2502 \u2514\u2500\u2500 Transition \u2502
\n\u2502 \u2502 \u2502
\n\u2502 \u251c\u2500\u2500 Pair Stack \u2502 Pair\u504f\u7f6e\u66f4\u65b0
\n\u2502 \u2502 \u251c\u2500\u2500 Triangle Multiplicative\u2502
\n\u2502 \u2502 \u251c\u2500\u2500 Triangle Self-Attention\u2502
\n\u2502 \u2502 \u2514\u2500\u2500 Transition \u2502
\n\u2502 \u2502 \u2502
\n\u2502 \u2514\u2500\u2500 MSA \u2194 Pair \u4fe1\u606f\u4ea4\u6362 \u2502 Outer Product Mean
\n\u2514\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u252c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2518
\n \u2502
\n \u25bc
\n\u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510
\n\u2502 Structure Module (8\u5c42) \u2502
\n\u2502 \u251c\u2500\u2500 IPA (Invariant Point \u2502
\n\u2502 \u2502 Attention) \u2502 SE(3)\u7b49\u53d8Transformer
\n\u2502 \u2502 \u2502
\n\u2502 \u2514\u2500\u2500 Backbone Update \u2502 \u9884\u6d4b\u65cb\u8f6c+\u5e73\u79fb
\n\u2502 (\u6bcf\u4e2a\u6b8b\u57fa\u7684\u521a\u4f53\u53d8\u6362) \u2502
\n\u2514\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u252c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2518
\n \u2502
\n \u25bc
\n 3D\u5750\u6807 (N, C\u03b1, C per residue)<\/p>\n

3.1 Evoformer\u6838\u5fc3：\u4e09\u89d2\u4e58\u6cd5\u66f4\u65b0<\/h4>\n

import<\/span> torch
\nimport<\/span> torch.<\/span>nn as<\/span> nn
\nimport<\/span> torch.<\/span>nn.<\/span>functional as<\/span> F<\/p>\n

class<\/span> TriangleMultiplication<\/span>(<\/span>nn.<\/span>Module)<\/span>:<\/span>
\n """\u4e09\u89d2\u4e58\u6cd5\u66f4\u65b0：\u5229\u7528Pair\u77e9\u9635\u7684\u51e0\u4f55\u7ea6\u675f
\n zij \u2190 zij + \u03a3_k z_ik \u00b7 z_kj (\u884c\u66f4\u65b0)
\n zij \u2190 zij + \u03a3_k z_ki \u00b7 z_jk (\u5217\u66f4\u65b0)
\n """<\/span><\/p>\n

def<\/span> __init__<\/span>(<\/span>self,<\/span> c_z=<\/span>128<\/span>,<\/span> c_hidden=<\/span>128<\/span>)<\/span>:<\/span>
\n super<\/span>(<\/span>)<\/span>.<\/span>__init__(<\/span>)<\/span>
\n self.<\/span>layer_norm_in =<\/span> nn.<\/span>LayerNorm(<\/span>c_z)<\/span>
\n self.<\/span>layer_norm_out =<\/span> nn.<\/span>LayerNorm(<\/span>c_z)<\/span><\/p>\n

# \u95e8\u63a7\u7ebf\u6027\u5c42<\/span>
\n self.<\/span>left_projection =<\/span> nn.<\/span>Linear(<\/span>c_z,<\/span> c_hidden)<\/span>
\n self.<\/span>right_projection =<\/span> nn.<\/span>Linear(<\/span>c_z,<\/span> c_hidden)<\/span>
\n self.<\/span>left_gate =<\/span> nn.<\/span>Linear(<\/span>c_z,<\/span> c_hidden)<\/span>
\n self.<\/span>right_gate =<\/span> nn.<\/span>Linear(<\/span>c_z,<\/span> c_hidden)<\/span><\/p>\n

self.<\/span>output_projection =<\/span> nn.<\/span>Linear(<\/span>c_hidden,<\/span> c_z)<\/span>
\n self.<\/span>gate =<\/span> nn.<\/span>Linear(<\/span>c_z,<\/span> c_z)<\/span><\/p>\n

def<\/span> forward<\/span>(<\/span>self,<\/span> z,<\/span> mask=<\/span>None<\/span>)<\/span>:<\/span>
\n """z: [B, N_res, N_res, c_z] Pair\u8868\u793a"""<\/span>
\n z =<\/span> self.<\/span>layer_norm_in(<\/span>z)<\/span><\/p>\n

# \u5de6\u6295\u5f71\u548c\u95e8\u63a7<\/span>
\n left_proj =<\/span> self.<\/span>left_projection(<\/span>z)<\/span> # [B,N,N,c_hidden]<\/span>
\n left_gate =<\/span> torch.<\/span>sigmoid(<\/span>self.<\/span>left_gate(<\/span>z)<\/span>)<\/span>
\n left =<\/span> left_proj *<\/span> left_gate<\/p>\n

# \u53f3\u6295\u5f71\u548c\u95e8\u63a7<\/span>
\n right_proj =<\/span> self.<\/span>right_projection(<\/span>z)<\/span>
\n right_gate =<\/span> torch.<\/span>sigmoid(<\/span>self.<\/span>right_gate(<\/span>z)<\/span>)<\/span>
\n right =<\/span> right_proj *<\/span> right_gate<\/p>\n

# \u884c\u66f4\u65b0: out_ij = \u03a3_k left_ik * right_jk<\/span>
\n # \u7b49\u4ef7\u4e8e out = left @ right^T（\u5728\u901a\u9053\u7ef4\u5ea6\u6c42\u548c）<\/span>
\n # \u5b9e\u73b0\u4e3aeinops\u98ce\u683c: b i k c, b j k c -> b i j c<\/span>
\n left =<\/span> left.<\/span>permute(<\/span>0<\/span>,<\/span> 3<\/span>,<\/span> 1<\/span>,<\/span> 2<\/span>)<\/span> # [B, c_hidden, N, N]<\/span>
\n right =<\/span> right.<\/span>permute(<\/span>0<\/span>,<\/span> 3<\/span>,<\/span> 2<\/span>,<\/span> 1<\/span>)<\/span> # [B, c_hidden, N, N]<\/span><\/p>\n

z_update =<\/span> torch.<\/span>einsum(<\/span>'bcij,bcjk->bcik'<\/span>,<\/span> left,<\/span> right)<\/span>
\n z_update =<\/span> z_update.<\/span>permute(<\/span>0<\/span>,<\/span> 2<\/span>,<\/span> 3<\/span>,<\/span> 1<\/span>)<\/span> # [B, N, N, c_hidden]<\/span><\/p>\n

# \u8f93\u51fa\u95e8\u63a7<\/span>
\n z_update =<\/span> self.<\/span>output_projection(<\/span>z_update)<\/span>
\n gate =<\/span> torch.<\/span>sigmoid(<\/span>self.<\/span>gate(<\/span>self.<\/span>layer_norm_out(<\/span>z)<\/span>)<\/span>)<\/span><\/p>\n

return<\/span> z +<\/span> gate *<\/span> z_update<\/p>\n

3.2 IPA（\u4e0d\u53d8\u70b9\u6ce8\u610f\u529b）<\/h4>\n

class<\/span> InvariantPointAttention<\/span>(<\/span>nn.<\/span>Module)<\/span>:<\/span>
\n """SE(3)\u7b49\u53d8\u7684\u4e0d\u53d8\u70b9\u6ce8\u610f\u529b
\n \u6838\u5fc3: \u75283D\u5750\u6807(r,t)\u6295\u5f71\u5230\u6ce8\u610f\u529b\u7684key\/query\u7a7a\u95f4
\n """<\/span><\/p>\n

def<\/span> __init__<\/span>(<\/span>self,<\/span> c_s=<\/span>384<\/span>,<\/span> c_z=<\/span>128<\/span>,<\/span> n_heads=<\/span>12<\/span>,<\/span> n_query_points=<\/span>4<\/span>,<\/span> n_value_points=<\/span>8<\/span>)<\/span>:<\/span>
\n super<\/span>(<\/span>)<\/span>.<\/span>__init__(<\/span>)<\/span>
\n self.<\/span>n_heads =<\/span> n_heads
\n self.<\/span>n_query_points =<\/span> n_query_points
\n self.<\/span>n_value_points =<\/span> n_value_points<\/p>\n

# \u7ebf\u6027\u6295\u5f71<\/span>
\n self.<\/span>linear_q =<\/span> nn.<\/span>Linear(<\/span>c_s,<\/span> c_s)<\/span>
\n self.<\/span>linear_kv =<\/span> nn.<\/span>Linear(<\/span>c_s,<\/span> 2<\/span> *<\/span> c_s)<\/span><\/p>\n

# \u70b9\u6295\u5f71（\u4e0d\u53d8\u6027\u6765\u81ea\u521a\u4f53\u53d8\u6362\u4e0b\u7684\u51e0\u4f55\u8ddd\u79bb）<\/span>
\n self.<\/span>linear_q_points =<\/span> nn.<\/span>Linear(<\/span>c_s,<\/span> n_heads *<\/span> n_query_points *<\/span> 3<\/span>)<\/span>
\n self.<\/span>linear_kv_points =<\/span> nn.<\/span>Linear(<\/span>c_s,<\/span> n_heads *<\/span> (<\/span>n_query_points +<\/span> n_value_points)<\/span> *<\/span> 3<\/span>)<\/span><\/p>\n

self.<\/span>linear_b =<\/span> nn.<\/span>Linear(<\/span>c_z,<\/span> n_heads)<\/span>
\n self.<\/span>linear_out =<\/span> nn.<\/span>Linear(<\/span>c_s,<\/span> c_s)<\/span><\/p>\n

def<\/span> forward<\/span>(<\/span>self,<\/span> s,<\/span> z,<\/span> rigids,<\/span> mask=<\/span>None<\/span>)<\/span>:<\/span>
\n """
\n s: [B, N, c_s] \u5355\u8868\u793a
\n z: [B, N, N, c_z] Pair\u8868\u793a
\n rigids: \u6bcf\u4e2a\u6b8b\u57fa\u7684\u521a\u4f53\u53d8\u6362(\u65cb\u8f6c\u77e9\u9635R + \u5e73\u79fbt)
\n """<\/span>
\n B,<\/span> N,<\/span> c_s =<\/span> s.<\/span>shape
\n H =<\/span> self.<\/span>n_heads
\n d_h =<\/span> c_s \/\/<\/span> H<\/p>\n

# 1. \u6807\u51c6\u6ce8\u610f\u529b\u6295\u5f71<\/span>
\n q =<\/span> self.<\/span>linear_q(<\/span>s)<\/span>.<\/span>reshape(<\/span>B,<\/span> N,<\/span> H,<\/span> d_h)<\/span> # [B,N,H,d_h]<\/span>
\n k,<\/span> v =<\/span> self.<\/span>linear_kv(<\/span>s)<\/span>.<\/span>chunk(<\/span>2<\/span>,<\/span> dim=<\/span>–<\/span>1<\/span>)<\/span>
\n k =<\/span> k.<\/span>reshape(<\/span>B,<\/span> N,<\/span> H,<\/span> d_h)<\/span>
\n v =<\/span> v.<\/span>reshape(<\/span>B,<\/span> N,<\/span> H,<\/span> d_h)<\/span><\/p>\n

# 2. \u4e0d\u53d8\u70b9\u6ce8\u610f\u529b\u7684\u70b9\u6295\u5f71<\/span>
\n # \u4e3a\u6bcf\u4e2a\u6b8b\u57fa\u751f\u62103D\u67e5\u8be2\u70b9\u548c\u5173\u952e\u70b9<\/span>
\n q_pts =<\/span> self.<\/span>linear_q_points(<\/span>s)<\/span>.<\/span>reshape(<\/span>B,<\/span> N,<\/span> H,<\/span> self.<\/span>n_query_points,<\/span> 3<\/span>)<\/span>
\n k_pts =<\/span> self.<\/span>linear_kv_points(<\/span>s)<\/span>.<\/span>reshape(<\/span>B,<\/span> N,<\/span> H,<\/span> self.<\/span>n_query_points +<\/span> self.<\/span>n_value_points,<\/span> 3<\/span>)<\/span>
\n k_pts,<\/span> v_pts =<\/span> torch.<\/span>split(<\/span>k_pts,<\/span> [<\/span>self.<\/span>n_query_points,<\/span> self.<\/span>n_value_points]<\/span>,<\/span> dim=<\/span>3<\/span>)<\/span><\/p>\n

# 3. \u5c06\u70b9\u6295\u5f71\u5230\u5168\u5c40\u5750\u6807（\u5e94\u7528\u521a\u4f53\u53d8\u6362 R\u00b7p + t）<\/span>
\n R,<\/span> t =<\/span> rigids # R: [B,N,3,3], t: [B,N,3]<\/span><\/p>\n

# \u5168\u5c40\u67e5\u8be2\u70b9<\/span>
\n q_pts_global =<\/span> torch.<\/span>einsum(<\/span>'bnrc,bnhpc->bnhpr'<\/span>,<\/span> R,<\/span> q_pts)<\/span> +<\/span> t[<\/span>:<\/span>,<\/span>:<\/span>,<\/span>None<\/span>,<\/span>None<\/span>,<\/span>:<\/span>]<\/span>
\n # \u5168\u5c40\u5173\u952e\u70b9<\/span>
\n k_pts_global =<\/span> torch.<\/span>einsum(<\/span>'bnrc,bnhpc->bnhpr'<\/span>,<\/span> R,<\/span> k_pts)<\/span> +<\/span> t[<\/span>:<\/span>,<\/span>:<\/span>,<\/span>None<\/span>,<\/span>None<\/span>,<\/span>:<\/span>]<\/span><\/p>\n

# 4. \u8ba1\u7b97\u57fa\u4e8e\u70b9\u8ddd\u79bb\u7684\u6ce8\u610f\u529b\u504f\u7f6e<\/span>
\n # dist_ij = ||q_i_p – k_j_p||\u00b2<\/span>
\n q_expand =<\/span> q_pts_global.<\/span>unsqueeze(<\/span>2<\/span>)<\/span> # [B,N,1,H,P_q,3]<\/span>
\n k_expand =<\/span> k_pts_global.<\/span>unsqueeze(<\/span>1<\/span>)<\/span> # [B,1,N,H,P_k,3]<\/span><\/p>\n

# \u6210\u5bf9\u5e73\u65b9\u8ddd\u79bb<\/span>
\n dist_sq =<\/span> (<\/span>(<\/span>q_expand –<\/span> k_expand)<\/span> **<\/span> 2<\/span>)<\/span>.<\/span>sum<\/span>(<\/span>dim=<\/span>–<\/span>1<\/span>)<\/span> # [B,N,N,H,P_q,P_k]<\/span>
\n point_bias =<\/span> –<\/span>0.5<\/span> *<\/span> dist_sq.<\/span>mean(<\/span>dim=<\/span>(<\/span>–<\/span>1<\/span>,<\/span> –<\/span>2<\/span>)<\/span>)<\/span> # \u5e73\u5747\u8ddd\u79bb\u4f5c\u4e3a\u504f\u7f6e<\/span><\/p>\n

# 5. \u6807\u51c6\u6ce8\u610f\u529b\u8ba1\u7b97<\/span>
\n attn_logits =<\/span> torch.<\/span>einsum(<\/span>'bihd,bjhd->bhij'<\/span>,<\/span> q,<\/span> k)<\/span> *<\/span> (<\/span>d_h **<\/span> –<\/span>0.5<\/span>)<\/span><\/p>\n

# \u52a0\u5165Pair\u504f\u7f6e<\/span>
\n pair_bias =<\/span> self.<\/span>linear_b(<\/span>z)<\/span>.<\/span>permute(<\/span>0<\/span>,<\/span> 3<\/span>,<\/span> 1<\/span>,<\/span> 2<\/span>)<\/span> # [B,H,N,N]<\/span>
\n attn_logits =<\/span> attn_logits +<\/span> pair_bias +<\/span> point_bias<\/p>\n

# \u8f6f\u6ce8\u610f\u529b<\/span>
\n if<\/span> mask is<\/span> not<\/span> None<\/span>:<\/span>
\n attn_logits =<\/span> attn_logits.<\/span>masked_fill(<\/span>~<\/span>mask[<\/span>:<\/span>,<\/span>None<\/span>,<\/span>None<\/span>,<\/span>:<\/span>]<\/span>,<\/span> –<\/span>1e9<\/span>)<\/span>
\n attn =<\/span> F.<\/span>softmax(<\/span>attn_logits,<\/span> dim=<\/span>–<\/span>1<\/span>)<\/span><\/p>\n

# 6. \u805a\u5408Value<\/span>
\n out =<\/span> torch.<\/span>einsum(<\/span>'bhij,bjhd->bihd'<\/span>,<\/span> attn,<\/span> v)<\/span>
\n out =<\/span> out.<\/span>reshape(<\/span>B,<\/span> N,<\/span> c_s)<\/span>
\n return<\/span> self.<\/span>linear_out(<\/span>out)<\/span><\/p>\n

\u56db\u3001\u5b8c\u6574\u63a8\u7406\u6d41\u7a0b<\/h3>\n

from<\/span> alphafold.<\/span>model import<\/span> model
\nfrom<\/span> alphafold.<\/span>data import<\/span> pipeline,<\/span> templates<\/p>\n

class<\/span> AlphaFoldInference<\/span>:<\/span>
\n def<\/span> __init__<\/span>(<\/span>self,<\/span> model_params_path,<\/span> database_path)<\/span>:<\/span>
\n # \u52a0\u8f7d\u6a21\u578b<\/span>
\n self.<\/span>model_config =<\/span> model.<\/span>config(<\/span>model_name=<\/span>"model_1"<\/span>)<\/span>
\n self.<\/span>model_params =<\/span> model.<\/span>get_model_haiku_params(<\/span>model_name=<\/span>"model_1"<\/span>,<\/span> data_dir=<\/span>model_params_path)<\/span>
\n self.<\/span>model_runner =<\/span> model.<\/span>RunModel(<\/span>self.<\/span>model_config,<\/span> self.<\/span>model_params)<\/span><\/p>\n

# \u6570\u636e\u7ba1\u9053<\/span>
\n self.<\/span>data_pipeline =<\/span> pipeline.<\/span>DataPipeline(<\/span>
\n jackhmmer_binary_path=<\/span>f"<\/span>{<\/span>database_path}<\/span><\/span>\/jackhmmer"<\/span><\/span>,<\/span>
\n hhblits_binary_path=<\/span>f"<\/span>{<\/span>database_path}<\/span><\/span>\/hhblits"<\/span><\/span>,<\/span>
\n uniref90_database_path=<\/span>f"<\/span>{<\/span>database_path}<\/span><\/span>\/uniref90"<\/span><\/span>,<\/span>
\n mgnify_database_path=<\/span>f"<\/span>{<\/span>database_path}<\/span><\/span>\/mgnify"<\/span><\/span>,<\/span>
\n bfd_database_path=<\/span>f"<\/span>{<\/span>database_path}<\/span><\/span>\/bfd"<\/span><\/span>,<\/span>
\n uniclust30_database_path=<\/span>f"<\/span>{<\/span>database_path}<\/span><\/span>\/uniclust30"<\/span><\/span>,<\/span>
\n pdb70_database_path=<\/span>f"<\/span>{<\/span>database_path}<\/span><\/span>\/pdb70"<\/span><\/span>,<\/span>
\n template_featurizer=<\/span>templates.<\/span>HhsearchHitFeaturizer(<\/span>.<\/span>.<\/span>.<\/span>)<\/span>,<\/span>
\n )<\/span><\/p>\n

def<\/span> predict<\/span>(<\/span>self,<\/span> sequence,<\/span> name=<\/span>"query"<\/span>)<\/span>:<\/span>
\n """\u7aef\u5230\u7aef\u9884\u6d4b"""<\/span>
\n features =<\/span> self.<\/span>data_pipeline.<\/span>process(<\/span>
\n input_fasta_path=<\/span>f"><\/span>{<\/span>name}<\/span><\/span>\\\\n<\/span>{<\/span>sequence}<\/span><\/span>"<\/span><\/span>,<\/span>
\n msa_output_dir=<\/span>"\/tmp\/msa"<\/span>
\n )<\/span><\/p>\n

# \u8fd0\u884c\u6a21\u578b（\u56de\u65363\u6b21）<\/span>
\n prediction =<\/span> self.<\/span>model_runner.<\/span>process_features(<\/span>
\n features,<\/span>
\n random_seed=<\/span>42<\/span>
\n )<\/span><\/p>\n

# \u63d0\u53d6\u7ed3\u679c<\/span>
\n result =<\/span> {<\/span>
\n "plddt"<\/span>:<\/span> prediction[<\/span>"plddt"<\/span>]<\/span>,<\/span> # \u9884\u6d4b\u7f6e\u4fe1\u5ea6<\/span>
\n "distogram"<\/span>:<\/span> prediction[<\/span>"distogram"<\/span>]<\/span>,<\/span>
\n "structure_module"<\/span>:<\/span> {<\/span>
\n "final_atom_positions"<\/span>:<\/span> prediction[<\/span>"structure_module"<\/span>]<\/span>[<\/span>"final_atom_positions"<\/span>]<\/span>,<\/span>
\n "final_atom_mask"<\/span>:<\/span> prediction[<\/span>"structure_module"<\/span>]<\/span>[<\/span>"final_atom_mask"<\/span>]<\/span>,<\/span>
\n }<\/span>,<\/span>
\n "predicted_aligned_error"<\/span>:<\/span> prediction.<\/span>get(<\/span>"predicted_aligned_error"<\/span>)<\/span>,<\/span>
\n }<\/span><\/p>\n

return<\/span> result<\/p>\n

def<\/span> save_pdb<\/span>(<\/span>self,<\/span> prediction,<\/span> output_path)<\/span>:<\/span>
\n """\u4fdd\u5b58\u4e3aPDB\u683c\u5f0f"""<\/span>
\n from<\/span> alphafold.<\/span>common import<\/span> protein<\/p>\n

plddt =<\/span> prediction[<\/span>"plddt"<\/span>]<\/span>
\n atom_positions =<\/span> prediction[<\/span>"structure_module"<\/span>]<\/span>[<\/span>"final_atom_positions"<\/span>]<\/span><\/p>\n

# \u8f6c\u4e3aProtein\u5bf9\u8c61<\/span>
\n unrelaxed_protein =<\/span> protein.<\/span>from_prediction(<\/span>
\n features=<\/span>None<\/span>,<\/span>
\n prediction=<\/span>prediction,<\/span>
\n )<\/span><\/p>\n

with<\/span> open<\/span>(<\/span>output_path,<\/span> 'w'<\/span>)<\/span> as<\/span> f:<\/span>
\n f.<\/span>write(<\/span>protein.<\/span>to_pdb(<\/span>unrelaxed_protein)<\/span>)<\/span><\/p>\n

print<\/span>(<\/span>f"PDB saved: <\/span>{<\/span>output_path}<\/span><\/span>"<\/span><\/span>)<\/span><\/p>\n

# \u4f7f\u7528<\/span>
\naf =<\/span> AlphaFoldInference(<\/span>".\/params"<\/span>,<\/span> ".\/databases"<\/span>)<\/span>
\nresult =<\/span> af.<\/span>predict(<\/span>"MKFLILFNILVCLAVAA…"<\/span>)<\/span>
\naf.<\/span>save_pdb(<\/span>result,<\/span> "prediction.pdb"<\/span>)<\/span>
\nprint<\/span>(<\/span>f"Mean pLDDT: <\/span>{<\/span>result[<\/span>'plddt'<\/span>]<\/span>.<\/span>mean(<\/span>)<\/span>:<\/span>.2f<\/span>}<\/span><\/span>"<\/span><\/span>)<\/span><\/p>\n

\u4e94\u3001AlphaFold3 \u65b0\u7279\u6027<\/h3>\n

# AlphaFold3\u5173\u952e\u6539\u8fdb:<\/span>
\n# 1. \u6269\u6563\u6a21\u5757\u66ff\u4ee3\u7ed3\u6784\u6a21\u5757 \u2192 \u76f4\u63a5\u751f\u6210\u539f\u5b50\u5750\u6807<\/span>
\n# 2. \u652f\u6301\u6240\u6709\u751f\u7269\u5206\u5b50（\u86cb\u767d\u8d28\/DNA\/RNA\/\u914d\u4f53\/\u79bb\u5b50）<\/span>
\n# 3. Pairformer\u66ff\u4ee3Evoformer \u2192 \u7b80\u5316MSA\u5904\u7406<\/span>
\n# 4. \u7f6e\u4fe1\u5ea6\u9884\u6d4b\u66f4\u51c6\u786e<\/span><\/p>\n

# \u4f7f\u7528AlphaFold3 (\u901a\u8fc7API)<\/span>
\nimport<\/span> requests<\/p>\n

def<\/span> predict_alphafold3<\/span>(<\/span>sequences)<\/span>:<\/span>
\n response =<\/span> requests.<\/span>post(<\/span>
\n "https:\/\/alphafold3-api.example.com\/predict"<\/span>,<\/span>
\n json=<\/span>{<\/span>
\n "sequences"<\/span>:<\/span> [<\/span>
\n {<\/span>"protein"<\/span>:<\/span> {<\/span>"sequence"<\/span>:<\/span> "MKFLIL…"<\/span>,<\/span> "id"<\/span>:<\/span> "A"<\/span>}<\/span>}<\/span>,<\/span>
\n {<\/span>"ligand"<\/span>:<\/span> {<\/span>"smiles"<\/span>:<\/span> "CC(=O)OC1=CC=CC=C1C(=O)O"<\/span>,<\/span> "id"<\/span>:<\/span> "B"<\/span>}<\/span>}<\/span>
\n ]<\/span>
\n }<\/span>
\n )<\/span>
\n return<\/span> response.<\/span>json(<\/span>)<\/span><\/p>\n

\u516d\u3001\u5173\u952e\u6307\u6807\u89e3\u8bfb<\/h3>\n\n\u6307\u6807\u542b\u4e49\u597d\u9608\u503c<\/tr>\n\n\n\n\n\n
pLDDT<\/td>\n\u9884\u6d4b\u5c40\u90e8\u8ddd\u79bb\u8bef\u5dee<\/td>\n>90 \u9ad8\u7f6e\u4fe1, <50 \u4f4e\u7f6e\u4fe1<\/td>\n<\/tr>\n
PAE<\/td>\n\u9884\u6d4b\u5bf9\u9f50\u8bef\u5dee<\/td>\n<5\u00c5 \u53ef\u4fe1, >15\u00c5 \u4e0d\u53ef\u9760<\/td>\n<\/tr>\n
pTM<\/td>\n\u9884\u6d4b\u6a21\u677f\u5efa\u6a21\u5206\u6570<\/td>\n>0.8 \u9ad8\u7f6e\u4fe1<\/td>\n<\/tr>\n
ipTM<\/td>\n\u754c\u9762pTM（\u590d\u5408\u4f53）<\/td>\n>0.8 \u9ad8\u7f6e\u4fe1<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

\u4e03\u3001\u603b\u7ed3<\/h3>\n

AlphaFold2\u7684\u6838\u5fc3\u8bbe\u8ba1：<\/p>\n

  • MSA + Pair \u2192 Evoformer \u2014 48\u5c42\u6df1\u5ea6\u5904\u7406\u5171\u8fdb\u5316\u4fe1\u606f<\/li>\n
  • IPA（\u4e0d\u53d8\u70b9\u6ce8\u610f\u529b） \u2014 SE(3)\u7b49\u53d8\u76843D\u7ed3\u6784\u751f\u6210<\/li>\n
  • \u56de\u6536\u673a\u5236 \u2014 3\u6b21\u8fed\u4ee3\u9010\u6b65\u7cbe\u5316\u9884\u6d4b<\/li>\n
  • \u7aef\u5230\u7aef\u8bad\u7ec3 \u2014 \u4ece\u5e8f\u5217\u76f4\u63a5\u9884\u6d4b3D\u5750\u6807<\/li>\n

    AlphaFold3\u5f15\u5165\u6269\u6563\u6a21\u578b\u548c\u5168\u539f\u5b50\u652f\u6301，\u5c06\u86cb\u767d\u8d28\u7ed3\u6784\u9884\u6d4b\u63a8\u5411\u4e86\u65b0\u7684\u9ad8\u5ea6\u3002<\/p>\n","protected":false},"excerpt":{"rendered":"

    AlphaFold\u539f\u7406\u4e0e\u86cb\u767d\u8d28\u7ed3\u6784\u9884\u6d4b\u5b9e\u6218
    \n\u4e00\u3001\u5f15\u8a00
    \n\\”\u86cb\u767d\u8d28\u6298\u53e0\u95ee\u9898\\”\u88ab\u79f0\u4e3a\u751f\u7269\u5b6650\u5e74\u672a\u89e3\u96be\u9898\u30022020\u5e74 DeepMind \u7684 AlphaFold2 \u5728 CASP14 \u4e2d\u4ee5\u539f\u5b50\u7ea7\u7cbe\u5ea6（\u4e2d\u4f4d\u8bef\u5dee0.96\u00c5）\u9884\u6d4b\u86cb\u767d\u8d283D\u7ed3\u6784，\u9707\u60ca\u79d1\u5b66\u754c\u30022024\u5e74 AlphaFold3 \u8fdb\u4e00\u6b65\u6269\u5c55\u5230\u6240\u6709\u751f\u7269\u5206\u5b50\u3002
    \n\u672c\u6587\u5c06\u6df1\u5165\u89e3\u6790 AlphaFold2 \u7684\u6838\u5fc3\u67b6\u6784：MSA\u5904\u7406\u3001Pair\u8868\u793a\u3001\u7ed3\u6784\u6a21\u5757（IPA&#xff09<\/p>\n","protected":false},"author":2,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1],"tags":[81,50,86],"topic":[],"class_list":["post-85614","post","type-post","status-publish","format-standard","hentry","category-server","tag-python","tag-50","tag-86"],"yoast_head":"\nAlphaFold\u539f\u7406\u4e0e\u86cb\u767d\u8d28\u7ed3\u6784\u9884\u6d4b\u5b9e\u6218 - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3<\/title>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/www.wsisp.com\/helps\/85614.html\" \/>\n<meta property=\"og:locale\" content=\"zh_CN\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"AlphaFold\u539f\u7406\u4e0e\u86cb\u767d\u8d28\u7ed3\u6784\u9884\u6d4b\u5b9e\u6218 - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3\" \/>\n<meta property=\"og:description\" content=\"AlphaFold\u539f\u7406\u4e0e\u86cb\u767d\u8d28\u7ed3\u6784\u9884\u6d4b\u5b9e\u6218 \u4e00\u3001\u5f15\u8a00 "\u86cb\u767d\u8d28\u6298\u53e0\u95ee\u9898"\u88ab\u79f0\u4e3a\u751f\u7269\u5b6650\u5e74\u672a\u89e3\u96be\u9898\u30022020\u5e74 DeepMind \u7684 AlphaFold2 \u5728 CASP14 \u4e2d\u4ee5\u539f\u5b50\u7ea7\u7cbe\u5ea6（\u4e2d\u4f4d\u8bef\u5dee0.96\u00c5）\u9884\u6d4b\u86cb\u767d\u8d283D\u7ed3\u6784，\u9707\u60ca\u79d1\u5b66\u754c\u30022024\u5e74 AlphaFold3 \u8fdb\u4e00\u6b65\u6269\u5c55\u5230\u6240\u6709\u751f\u7269\u5206\u5b50\u3002 \u672c\u6587\u5c06\u6df1\u5165\u89e3\u6790 AlphaFold2 \u7684\u6838\u5fc3\u67b6\u6784：MSA\u5904\u7406\u3001Pair\u8868\u793a\u3001\u7ed3\u6784\u6a21\u5757（IPA&#xff09\" \/>\n<meta property=\"og:url\" content=\"https:\/\/www.wsisp.com\/helps\/85614.html\" \/>\n<meta property=\"og:site_name\" content=\"\u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3\" \/>\n<meta property=\"article:published_time\" content=\"2026-07-27T11:08:22+00:00\" \/>\n<meta name=\"author\" content=\"admin\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:label1\" content=\"\u4f5c\u8005\" \/>\n\t<meta name=\"twitter:data1\" content=\"admin\" \/>\n\t<meta name=\"twitter:label2\" content=\"\u9884\u8ba1\u9605\u8bfb\u65f6\u95f4\" \/>\n\t<meta name=\"twitter:data2\" content=\"6 \u5206\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\/\/schema.org\",\"@graph\":[{\"@type\":\"WebPage\",\"@id\":\"https:\/\/www.wsisp.com\/helps\/85614.html\",\"url\":\"https:\/\/www.wsisp.com\/helps\/85614.html\",\"name\":\"AlphaFold\u539f\u7406\u4e0e\u86cb\u767d\u8d28\u7ed3\u6784\u9884\u6d4b\u5b9e\u6218 - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3\",\"isPartOf\":{\"@id\":\"https:\/\/www.wsisp.com\/helps\/#website\"},\"datePublished\":\"2026-07-27T11:08:22+00:00\",\"dateModified\":\"2026-07-27T11:08:22+00:00\",\"author\":{\"@id\":\"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/358e386c577a3ab51c4493330a20ad41\"},\"breadcrumb\":{\"@id\":\"https:\/\/www.wsisp.com\/helps\/85614.html#breadcrumb\"},\"inLanguage\":\"zh-Hans\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\/\/www.wsisp.com\/helps\/85614.html\"]}]},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\/\/www.wsisp.com\/helps\/85614.html#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"\u9996\u9875\",\"item\":\"https:\/\/www.wsisp.com\/helps\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"AlphaFold\u539f\u7406\u4e0e\u86cb\u767d\u8d28\u7ed3\u6784\u9884\u6d4b\u5b9e\u6218\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\/\/www.wsisp.com\/helps\/#website\",\"url\":\"https:\/\/www.wsisp.com\/helps\/\",\"name\":\"\u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3\",\"description\":\"\u9999\u6e2f\u670d\u52a1\u5668_\u9999\u6e2f\u4e91\u670d\u52a1\u5668\u8d44\u8baf_\u670d\u52a1\u5668\u5e2e\u52a9\u6587\u6863_\u670d\u52a1\u5668\u6559\u7a0b\",\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\/\/www.wsisp.com\/helps\/?s={search_term_string}\"},\"query-input\":\"required name=search_term_string\"}],\"inLanguage\":\"zh-Hans\"},{\"@type\":\"Person\",\"@id\":\"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/358e386c577a3ab51c4493330a20ad41\",\"name\":\"admin\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"zh-Hans\",\"@id\":\"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/image\/\",\"url\":\"https:\/\/gravatar.wp-china-yes.net\/avatar\/?s=96&d=mystery\",\"contentUrl\":\"https:\/\/gravatar.wp-china-yes.net\/avatar\/?s=96&d=mystery\",\"caption\":\"admin\"},\"sameAs\":[\"http:\/\/wp.wsisp.com\"],\"url\":\"https:\/\/www.wsisp.com\/helps\/author\/admin\"}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"AlphaFold\u539f\u7406\u4e0e\u86cb\u767d\u8d28\u7ed3\u6784\u9884\u6d4b\u5b9e\u6218 - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/www.wsisp.com\/helps\/85614.html","og_locale":"zh_CN","og_type":"article","og_title":"AlphaFold\u539f\u7406\u4e0e\u86cb\u767d\u8d28\u7ed3\u6784\u9884\u6d4b\u5b9e\u6218 - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3","og_description":"AlphaFold\u539f\u7406\u4e0e\u86cb\u767d\u8d28\u7ed3\u6784\u9884\u6d4b\u5b9e\u6218 \u4e00\u3001\u5f15\u8a00 \"\u86cb\u767d\u8d28\u6298\u53e0\u95ee\u9898\"\u88ab\u79f0\u4e3a\u751f\u7269\u5b6650\u5e74\u672a\u89e3\u96be\u9898\u30022020\u5e74 DeepMind \u7684 AlphaFold2 \u5728 CASP14 \u4e2d\u4ee5\u539f\u5b50\u7ea7\u7cbe\u5ea6（\u4e2d\u4f4d\u8bef\u5dee0.96\u00c5）\u9884\u6d4b\u86cb\u767d\u8d283D\u7ed3\u6784，\u9707\u60ca\u79d1\u5b66\u754c\u30022024\u5e74 AlphaFold3 \u8fdb\u4e00\u6b65\u6269\u5c55\u5230\u6240\u6709\u751f\u7269\u5206\u5b50\u3002 \u672c\u6587\u5c06\u6df1\u5165\u89e3\u6790 AlphaFold2 \u7684\u6838\u5fc3\u67b6\u6784：MSA\u5904\u7406\u3001Pair\u8868\u793a\u3001\u7ed3\u6784\u6a21\u5757（IPA&#xff09","og_url":"https:\/\/www.wsisp.com\/helps\/85614.html","og_site_name":"\u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3","article_published_time":"2026-07-27T11:08:22+00:00","author":"admin","twitter_card":"summary_large_image","twitter_misc":{"\u4f5c\u8005":"admin","\u9884\u8ba1\u9605\u8bfb\u65f6\u95f4":"6 \u5206"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"WebPage","@id":"https:\/\/www.wsisp.com\/helps\/85614.html","url":"https:\/\/www.wsisp.com\/helps\/85614.html","name":"AlphaFold\u539f\u7406\u4e0e\u86cb\u767d\u8d28\u7ed3\u6784\u9884\u6d4b\u5b9e\u6218 - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3","isPartOf":{"@id":"https:\/\/www.wsisp.com\/helps\/#website"},"datePublished":"2026-07-27T11:08:22+00:00","dateModified":"2026-07-27T11:08:22+00:00","author":{"@id":"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/358e386c577a3ab51c4493330a20ad41"},"breadcrumb":{"@id":"https:\/\/www.wsisp.com\/helps\/85614.html#breadcrumb"},"inLanguage":"zh-Hans","potentialAction":[{"@type":"ReadAction","target":["https:\/\/www.wsisp.com\/helps\/85614.html"]}]},{"@type":"BreadcrumbList","@id":"https:\/\/www.wsisp.com\/helps\/85614.html#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"\u9996\u9875","item":"https:\/\/www.wsisp.com\/helps"},{"@type":"ListItem","position":2,"name":"AlphaFold\u539f\u7406\u4e0e\u86cb\u767d\u8d28\u7ed3\u6784\u9884\u6d4b\u5b9e\u6218"}]},{"@type":"WebSite","@id":"https:\/\/www.wsisp.com\/helps\/#website","url":"https:\/\/www.wsisp.com\/helps\/","name":"\u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3","description":"\u9999\u6e2f\u670d\u52a1\u5668_\u9999\u6e2f\u4e91\u670d\u52a1\u5668\u8d44\u8baf_\u670d\u52a1\u5668\u5e2e\u52a9\u6587\u6863_\u670d\u52a1\u5668\u6559\u7a0b","potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/www.wsisp.com\/helps\/?s={search_term_string}"},"query-input":"required name=search_term_string"}],"inLanguage":"zh-Hans"},{"@type":"Person","@id":"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/358e386c577a3ab51c4493330a20ad41","name":"admin","image":{"@type":"ImageObject","inLanguage":"zh-Hans","@id":"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/image\/","url":"https:\/\/gravatar.wp-china-yes.net\/avatar\/?s=96&d=mystery","contentUrl":"https:\/\/gravatar.wp-china-yes.net\/avatar\/?s=96&d=mystery","caption":"admin"},"sameAs":["http:\/\/wp.wsisp.com"],"url":"https:\/\/www.wsisp.com\/helps\/author\/admin"}]}},"_links":{"self":[{"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/posts\/85614","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/comments?post=85614"}],"version-history":[{"count":0,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/posts\/85614\/revisions"}],"wp:attachment":[{"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/media?parent=85614"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/categories?post=85614"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/tags?post=85614"},{"taxonomy":"topic","embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/topic?post=85614"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}