An introduction for microbiologists: drawings, questions, what we learned, glossary, tour

build.json

@@ -1,6 +1,6 @@
 {
- "git_sha": "6855fe324c52",
+ "git_sha": "33f70ff2b875",
  "dirty": false,
- "assembled_at": "2026-09-27T09:32:53Z",
- "page_md5": "b10af8db4056664578a25abfcc70b2ed"
+ "assembled_at": "2026-09-27T17:22:33Z",
+ "page_md5": "4a0bd360600dc918a351826f2b1779e2"
 }

pgb/readings/G8B09_RS13610.json

@@ -0,0 +1,427 @@
+{
+ "status": "llm",
+ "model": "deepseek-ai/DeepSeek-V4.1-Flash",
+ "provider": "deepinfra",
+ "created": 1790526995.984362,
+ "usage": {
+  "prompt_tokens": 12099,
+  "completion_tokens": 822,
+  "cost_usd": 0.001223,
+  "ms": 13245
+ },
+ "attempts": [
+  {
+   "text": "**The window opens on a short spine.** Columns 0 to 2, `yicS` to `yicL`, are carried by 97 to 100% of chromosomes, all persistent, in a region of plasticity in 0.2 to 2% of calls. The model names 95 to 99% of them, fewest at column 1, `nlpA`, 95%.\n\n**The model names more genes on the spine columns than on the fork and variable ones**, 85 against 82%, and hesitates more there, 0.71 against 0.41 bits. It hesitates most at column 4, `yicJ`, a spine column, 2.66 bits.\n\n**The window ends in variable columns.** Columns 67 to 79, `yibF` to `ulaD`, are carried by 7 to 62%. The last column is `ulaD`, in 18% of chromosomes.\n\n**Across the window it names 83% of genes**, 0.52 bits on average, perplexity 2.2 families. Inside a region of plasticity it hesitates more than outside, 1.22 against 0.20 bits, and names fewer genes, 61% against 93%.",
+   "finish": "stop",
+   "usage": {
+    "prompt_tokens": 5676,
+    "completion_tokens": 414,
+    "cached_tokens": 0,
+    "reasoning_tokens": 0,
+    "cost_usd": 0.000969
+   },
+   "provider": "deepinfra",
+   "ms": 6696,
+   "raw": "@start **The window opens on a short spine.** Columns 0 to 2, `yicS` to `yicL`, are carried by 97 to 100% of chromosomes, all persistent, in a region of plasticity in 0.2 to 2% of calls {S1, P}. The model names 95 to 99% of them, fewest at column 1, `nlpA`, 95% {S1}.\n\n@variable **A fork follows at `setC`**, column 3, carried by 73%, with a second family in 32% {S2.c3}. Then columns 4 to 15, `yicJ` to `yicG`, are the stretch most in regions of plasticity, 29 to 50%, [[lipopolysaccharide core]] {S3, P, Q}. Through columns 32 to 36 a branch of 5 shell families runs in 16%, [[lipopolysaccharide]] {S8.b32, Q}.\n\n@model **The model names more genes on the spine columns than on the fork and variable ones**, 85 against 82%, and hesitates more there, 0.71 against 0.41 bits {M}. It hesitates most at column 4, `yicJ`, a spine column, 2.66 bits {H}.\n\n@end **The window ends in variable columns.** Columns 67 to 79, `yibF` to `ulaD`, are carried by 7 to 62% {S8.c67-79}. The last column is `ulaD`, in 18% of chromosomes {E}.\n\n@whole **Across the window it names 83% of genes**, 0.52 bits on average, perplexity 2.2 families {G}. Inside a region of plasticity it hesitates more than outside, 1.22 against 0.20 bits, and names fewer genes, 61% against 93% {C}.",
+   "annotated": "@start **The window opens on a short spine.** Columns 0 to 2, `yicS` to `yicL`, are carried by 97 to 100% of chromosomes, all persistent, in a region of plasticity in 0.2 to 2% of calls {S1, P}. The model names 95 to 99% of them, fewest at column 1, `nlpA`, 95% {S1}.\n\n@variable **A fork follows at `setC`**, column 3, carried by 73%, with a second family in 32% {S2.c3}. Then columns 4 to 15, `yicJ` to `yicG`, are the stretch most in regions of plasticity, 29 to 50%, [[lipopolysaccharide core]] {S3, P, Q}. Through columns 32 to 36 a branch of 5 shell families runs in 16%, [[lipopolysaccharide]] {S8.b32, Q}.\n\n@model **The model names more genes on the spine columns than on the fork and variable ones**, 85 against 82%, and hesitates more there, 0.71 against 0.41 bits {M}. It hesitates most at column 4, `yicJ`, a spine column, 2.66 bits {H}.\n\n@end **The window ends in variable columns.** Columns 67 to 79, `yibF` to `ulaD`, are carried by 7 to 62% {S8.c67-79}. The last column is `ulaD`, in 18% of chromosomes {E}.\n\n@whole **Across the window it names 83% of genes**, 0.52 bits on average, perplexity 2.2 families {G}. Inside a region of plasticity it hesitates more than outside, 1.22 against 0.20 bits, and names fewer genes, 61% against 93% {C}.",
+   "check": {
+    "ok": false,
+    "notes": {
+     "tail": [],
+     "untagged": "",
+     "dup": [],
+     "tags": [
+      "start",
+      "variable",
+      "model",
+      "end",
+      "whole"
+     ],
+     "order": [
+      "start",
+      "variable",
+      "model",
+      "end",
+      "whole"
+     ]
+    },
+    "bad": [
+     {
+      "kind": "number",
+      "text": "32%",
+      "why": "column 3 is carried by 73% in the lines cited (S2.c3)",
+      "context": "**A fork follows at `setC`**, column 3, carried by 73%, with a second family in 32%"
+     },
+     {
+      "kind": "interpretation",
+      "text": "lipopolysaccharide core",
+      "why": "this says what 0 of the 12 families of the lines cited (S3, P, Q) do: name the columns it is true of, or leave the stretch unnamed",
+      "context": "[[lipopolysaccharide core]]"
+     },
+     {
+      "kind": "interpretation",
+      "text": "lipopolysaccharide",
+      "why": "an interpretation is one phrase of 2 to 6 words, not a word or a list",
+      "context": "[[lipopolysaccharide]]"
+     }
+    ],
+    "numbers": 27,
+    "names": 10,
+    "words": 204,
+    "sentences": 11,
+    "cited": 15,
+    "structure": []
+   },
+   "report": {
+    "dropped": 3,
+    "dropped_len": 0,
+    "unled": 0,
+    "sections": [
+     "start",
+     "model",
+     "end",
+     "whole"
+    ],
+    "words": 152,
+    "repeats": 0,
+    "interp": 0,
+    "kept": [
+     {
+      "tag": "start",
+      "text": "**The window opens on a short spine.** Columns 0 to 2, `yicS` to `yicL`, are carried by 97 to 100% of chromosomes, all persistent, in a region of plasticity in 0.2 to 2% of calls",
+      "cites": [
+       "S1",
+       "P"
+      ],
+      "inherit": []
+     },
+     {
+      "tag": "start",
+      "text": "The model names 95 to 99% of them, fewest at column 1, `nlpA`, 95%",
+      "cites": [
+       "S1"
+      ],
+      "inherit": []
+     },
+     {
+      "tag": "model",
+      "text": "**The model names more genes on the spine columns than on the fork and variable ones**, 85 against 82%, and hesitates more there, 0.71 against 0.41 bits",
+      "cites": [
+       "M"
+      ],
+      "inherit": [
+       "H"
+      ]
+     },
+     {
+      "tag": "model",
+      "text": "It hesitates most at column 4, `yicJ`, a spine column, 2.66 bits",
+      "cites": [
+       "H"
+      ],
+      "inherit": []
+     },
+     {
+      "tag": "end",
+      "text": "**The window ends in variable columns.** Columns 67 to 79, `yibF` to `ulaD`, are carried by 7 to 62%",
+      "cites": [
+       "S8.c67-79"
+      ],
+      "inherit": [
+       "E"
+      ]
+     },
+     {
+      "tag": "end",
+      "text": "The last column is `ulaD`, in 18% of chromosomes",
+      "cites": [
+       "E"
+      ],
+      "inherit": []
+     },
+     {
+      "tag": "whole",
+      "text": "**Across the window it names 83% of genes**, 0.52 bits on average, perplexity 2.2 families",
+      "cites": [
+       "G"
+      ],
+      "inherit": [
+       "C"
+      ]
+     },
+     {
+      "tag": "whole",
+      "text": "Inside a region of plasticity it hesitates more than outside, 1.22 against 0.20 bits, and names fewer genes, 61% against 93%",
+      "cites": [
+       "C"
+      ],
+      "inherit": []
+     }
+    ],
+    "covered": 17
+   },
+   "dropped": [
+    {
+     "tag": "variable",
+     "text": "**A fork follows at `setC`**, column 3, carried by 73%, with a second family in 32%",
+     "cites": [
+      "S2.c3"
+     ],
+     "opens": true,
+     "bad": [
+      {
+       "kind": "number",
+       "text": "32%",
+       "why": "column 3 is carried by 73% in the lines cited (S2.c3)"
+      }
+     ]
+    },
+    {
+     "tag": "variable",
+     "text": "Then columns 4 to 15, `yicJ` to `yicG`, are the stretch most in regions of plasticity, 29 to 50%, [[lipopolysaccharide core]]",
+     "cites": [
+      "S3",
+      "P",
+      "Q"
+     ],
+     "opens": false,
+     "bad": [
+      {
+       "kind": "interpretation",
+       "text": "lipopolysaccharide core",
+       "why": "this says what 0 of the 12 families of the lines cited (S3, P, Q) do: name the columns it is true of, or leave the stretch unnamed"
+      }
+     ]
+    },
+    {
+     "tag": "variable",
+     "text": "Through columns 32 to 36 a branch of 5 shell families runs in 16%, [[lipopolysaccharide]]",
+     "cites": [
+      "S8.b32",
+      "Q"
+     ],
+     "opens": false,
+     "bad": [
+      {
+       "kind": "interpretation",
+       "text": "lipopolysaccharide",
+       "why": "an interpretation is one phrase of 2 to 6 words, not a word or a list"
+      }
+     ]
+    }
+   ],
+   "served": "**The window opens on a short spine.** Columns 0 to 2, `yicS` to `yicL`, are carried by 97 to 100% of chromosomes, all persistent, in a region of plasticity in 0.2 to 2% of calls. The model names 95 to 99% of them, fewest at column 1, `nlpA`, 95%.\n\n**The model names more genes on the spine columns than on the fork and variable ones**, 85 against 82%, and hesitates more there, 0.71 against 0.41 bits. It hesitates most at column 4, `yicJ`, a spine column, 2.66 bits.\n\n**The window ends in variable columns.** Columns 67 to 79, `yibF` to `ulaD`, are carried by 7 to 62%. The last column is `ulaD`, in 18% of chromosomes.\n\n**Across the window it names 83% of genes**, 0.52 bits on average, perplexity 2.2 families. Inside a region of plasticity it hesitates more than outside, 1.22 against 0.20 bits, and names fewer genes, 61% against 93%.",
+   "truncated": false,
+   "ok": false
+  },
+  {
+   "text": "**The window opens on a short spine.** Columns 0 to 2, `yicS` to `yicL`, are carried by 97 to 100% of chromosomes, all persistent, in a region of plasticity in 0.2 to 2% of calls. The model names 95 to 99% of them, fewest at column 1, `nlpA`, 95%.\n\n**Then columns 4 to 15**, `yicJ` to `yicG`, are the stretch most in regions of plasticity, 29 to 50%. Through columns 32 to 36 a branch of 5 shell families runs in 16%, [[lipopolysaccharide core biosynthesis]].\n\n**The model names more genes on the spine columns than on the fork and variable ones**, 85 against 82%, and hesitates more there, 0.71 against 0.41 bits. It hesitates most at column 4, `yicJ`, a spine column, 2.66 bits.\n\n**The window ends in variable columns.** Columns 67 to 79, `yibF` to `ulaD`, are carried by 7 to 62%. The last column is `ulaD`, in 18% of chromosomes.\n\n**Across the window it names 83% of genes**, 0.52 bits on average, perplexity 2.2 families. Inside a region of plasticity it hesitates more than outside, 1.22 against 0.20 bits, and names fewer genes, 61% against 93%.",
+   "finish": "stop",
+   "usage": {
+    "prompt_tokens": 6423,
+    "completion_tokens": 408,
+    "cached_tokens": 6016,
+    "reasoning_tokens": 0,
+    "cost_usd": 0.000254
+   },
+   "provider": "deepinfra",
+   "ms": 6549,
+   "raw": "@start **The window opens on a short spine.** Columns 0 to 2, `yicS` to `yicL`, are carried by 97 to 100% of chromosomes, all persistent, in a region of plasticity in 0.2 to 2% of calls {S1, P}. The model names 95 to 99% of them, fewest at column 1, `nlpA`, 95% {S1}.\n\n@variable **A fork follows at `setC`**, column 3, carried by 73%, with a second family in 32% {S2.c3}. Then columns 4 to 15, `yicJ` to `yicG`, are the stretch most in regions of plasticity, 29 to 50% {S3, P}. Through columns 32 to 36 a branch of 5 shell families runs in 16%, [[lipopolysaccharide core biosynthesis]] {S8.b32, Q}.\n\n@model **The model names more genes on the spine columns than on the fork and variable ones**, 85 against 82%, and hesitates more there, 0.71 against 0.41 bits {M}. It hesitates most at column 4, `yicJ`, a spine column, 2.66 bits {H}.\n\n@end **The window ends in variable columns.** Columns 67 to 79, `yibF` to `ulaD`, are carried by 7 to 62% {S8.c67-79}. The last column is `ulaD`, in 18% of chromosomes {E}.\n\n@whole **Across the window it names 83% of genes**, 0.52 bits on average, perplexity 2.2 families {G}. Inside a region of plasticity it hesitates more than outside, 1.22 against 0.20 bits, and names fewer genes, 61% against 93% {C}.",
+   "annotated": "@start **The window opens on a short spine.** Columns 0 to 2, `yicS` to `yicL`, are carried by 97 to 100% of chromosomes, all persistent, in a region of plasticity in 0.2 to 2% of calls {S1, P}. The model names 95 to 99% of them, fewest at column 1, `nlpA`, 95% {S1}.\n\n@variable **A fork follows at `setC`**, column 3, carried by 73%, with a second family in 32% {S2.c3}. Then columns 4 to 15, `yicJ` to `yicG`, are the stretch most in regions of plasticity, 29 to 50% {S3, P}. Through columns 32 to 36 a branch of 5 shell families runs in 16%, [[lipopolysaccharide core biosynthesis]] {S8.b32, Q}.\n\n@model **The model names more genes on the spine columns than on the fork and variable ones**, 85 against 82%, and hesitates more there, 0.71 against 0.41 bits {M}. It hesitates most at column 4, `yicJ`, a spine column, 2.66 bits {H}.\n\n@end **The window ends in variable columns.** Columns 67 to 79, `yibF` to `ulaD`, are carried by 7 to 62% {S8.c67-79}. The last column is `ulaD`, in 18% of chromosomes {E}.\n\n@whole **Across the window it names 83% of genes**, 0.52 bits on average, perplexity 2.2 families {G}. Inside a region of plasticity it hesitates more than outside, 1.22 against 0.20 bits, and names fewer genes, 61% against 93% {C}.",
+   "check": {
+    "ok": false,
+    "notes": {
+     "tail": [],
+     "untagged": "",
+     "dup": [],
+     "tags": [
+      "start",
+      "variable",
+      "model",
+      "end",
+      "whole"
+     ],
+     "order": [
+      "start",
+      "variable",
+      "model",
+      "end",
+      "whole"
+     ]
+    },
+    "bad": [
+     {
+      "kind": "number",
+      "text": "32%",
+      "why": "column 3 is carried by 73% in the lines cited (S2.c3)",
+      "context": "**A fork follows at `setC`**, column 3, carried by 73%, with a second family in 32%"
+     }
+    ],
+    "numbers": 27,
+    "names": 10,
+    "words": 204,
+    "sentences": 11,
+    "cited": 14,
+    "structure": []
+   },
+   "report": {
+    "dropped": 1,
+    "dropped_len": 0,
+    "unled": 1,
+    "sections": [
+     "start",
+     "variable",
+     "model",
+     "end",
+     "whole"
+    ],
+    "words": 188,
+    "repeats": 0,
+    "interp": 1,
+    "kept": [
+     {
+      "tag": "start",
+      "text": "**The window opens on a short spine.** Columns 0 to 2, `yicS` to `yicL`, are carried by 97 to 100% of chromosomes, all persistent, in a region of plasticity in 0.2 to 2% of calls",
+      "cites": [
+       "S1",
+       "P"
+      ],
+      "inherit": []
+     },
+     {
+      "tag": "start",
+      "text": "The model names 95 to 99% of them, fewest at column 1, `nlpA`, 95%",
+      "cites": [
+       "S1"
+      ],
+      "inherit": []
+     },
+     {
+      "tag": "variable",
+      "text": "Then columns 4 to 15, `yicJ` to `yicG`, are the stretch most in regions of plasticity, 29 to 50%",
+      "cites": [
+       "S3",
+       "P"
+      ],
+      "inherit": []
+     },
+     {
+      "tag": "variable",
+      "text": "Through columns 32 to 36 a branch of 5 shell families runs in 16%, [[lipopolysaccharide core biosynthesis]]",
+      "cites": [
+       "S8.b32",
+       "Q"
+      ],
+      "inherit": []
+     },
+     {
+      "tag": "model",
+      "text": "**The model names more genes on the spine columns than on the fork and variable ones**, 85 against 82%, and hesitates more there, 0.71 against 0.41 bits",
+      "cites": [
+       "M"
+      ],
+      "inherit": [
+       "H"
+      ]
+     },
+     {
+      "tag": "model",
+      "text": "It hesitates most at column 4, `yicJ`, a spine column, 2.66 bits",
+      "cites": [
+       "H"
+      ],
+      "inherit": []
+     },
+     {
+      "tag": "end",
+      "text": "**The window ends in variable columns.** Columns 67 to 79, `yibF` to `ulaD`, are carried by 7 to 62%",
+      "cites": [
+       "S8.c67-79"
+      ],
+      "inherit": [
+       "E"
+      ]
+     },
+     {
+      "tag": "end",
+      "text": "The last column is `ulaD`, in 18% of chromosomes",
+      "cites": [
+       "E"
+      ],
+      "inherit": []
+     },
+     {
+      "tag": "whole",
+      "text": "**Across the window it names 83% of genes**, 0.52 bits on average, perplexity 2.2 families",
+      "cites": [
+       "G"
+      ],
+      "inherit": [
+       "C"
+      ]
+     },
+     {
+      "tag": "whole",
+      "text": "Inside a region of plasticity it hesitates more than outside, 1.22 against 0.20 bits, and names fewer genes, 61% against 93%",
+      "cites": [
+       "C"
+      ],
+      "inherit": []
+     }
+    ],
+    "covered": 33
+   },
+   "dropped": [
+    {
+     "tag": "variable",
+     "text": "**A fork follows at `setC`**, column 3, carried by 73%, with a second family in 32%",
+     "cites": [
+      "S2.c3"
+     ],
+     "opens": true,
+     "bad": [
+      {
+       "kind": "number",
+       "text": "32%",
+       "why": "column 3 is carried by 73% in the lines cited (S2.c3)"
+      }
+     ]
+    }
+   ],
+   "served": "**The window opens on a short spine.** Columns 0 to 2, `yicS` to `yicL`, are carried by 97 to 100% of chromosomes, all persistent, in a region of plasticity in 0.2 to 2% of calls. The model names 95 to 99% of them, fewest at column 1, `nlpA`, 95%.\n\n**Then columns 4 to 15**, `yicJ` to `yicG`, are the stretch most in regions of plasticity, 29 to 50%. Through columns 32 to 36 a branch of 5 shell families runs in 16%, [[lipopolysaccharide core biosynthesis]].\n\n**The model names more genes on the spine columns than on the fork and variable ones**, 85 against 82%, and hesitates more there, 0.71 against 0.41 bits. It hesitates most at column 4, `yicJ`, a spine column, 2.66 bits.\n\n**The window ends in variable columns.** Columns 67 to 79, `yibF` to `ulaD`, are carried by 7 to 62%. The last column is `ulaD`, in 18% of chromosomes.\n\n**Across the window it names 83% of genes**, 0.52 bits on average, perplexity 2.2 families. Inside a region of plasticity it hesitates more than outside, 1.22 against 0.20 bits, and names fewer genes, 61% against 93%.",
+   "truncated": false,
+   "ok": false
+  }
+ ],
+ "checked": {
+  "ok": true,
+  "numbers": 27,
+  "names": 10,
+  "words": 188,
+  "attempts": 2,
+  "sentences": 11,
+  "cited": 14,
+  "skill": "4a7fdc464826",
+  "dropped": 1,
+  "trimmed": 0,
+  "repeats": 0,
+  "unled": 1,
+  "sections": [
+   "start",
+   "variable",
+   "model",
+   "end",
+   "whole"
+  ],
+  "interp": 1,
+  "covered": 33
+ },
+ "text": "**The window opens on a short spine.** Columns 0 to 2, `yicS` to `yicL`, are carried by 97 to 100% of chromosomes, all persistent, in a region of plasticity in 0.2 to 2% of calls. The model names 95 to 99% of them, fewest at column 1, `nlpA`, 95%.\n\n**Then columns 4 to 15**, `yicJ` to `yicG`, are the stretch most in regions of plasticity, 29 to 50%. Through columns 32 to 36 a branch of 5 shell families runs in 16%, [[lipopolysaccharide core biosynthesis]].\n\n**The model names more genes on the spine columns than on the fork and variable ones**, 85 against 82%, and hesitates more there, 0.71 against 0.41 bits. It hesitates most at column 4, `yicJ`, a spine column, 2.66 bits.\n\n**The window ends in variable columns.** Columns 67 to 79, `yibF` to `ulaD`, are carried by 7 to 62%. The last column is `ulaD`, in 18% of chromosomes.\n\n**Across the window it names 83% of genes**, 0.52 bits on average, perplexity 2.2 families. Inside a region of plasticity it hesitates more than outside, 1.22 against 0.20 bits, and names fewer genes, 61% against 93%.",
+ "html": "<p class=\"note\"><b>The window opens on a short spine.</b> Columns 0 to 2, <span class=\"gene\">yicS</span> to <span class=\"gene\">yicL</span>, are carried by 97 to 100% of chromosomes, all persistent, in a region of plasticity in 0.2 to 2% of calls. The model names 95 to 99% of them, fewest at column 1, <span class=\"gene\">nlpA</span>, 95%.</p><p class=\"note\"><b>Then columns 4 to 15</b>, <span class=\"gene\">yicJ</span> to <span class=\"gene\">yicG</span>, are the stretch most in regions of plasticity, 29 to 50%. Through columns 32 to 36 a branch of 5 shell families runs in 16%, <span class=\"interp\" title=\"interpretation from the gene names and products, not measured here\">lipopolysaccharide core biosynthesis</span>.</p><p class=\"note\"><b>The model names more genes on the spine columns than on the fork and variable ones</b>, 85 against 82%, and hesitates more there, 0.71 against 0.41 bits. It hesitates most at column 4, <span class=\"gene\">yicJ</span>, a spine column, 2.66 bits.</p><p class=\"note\"><b>The window ends in variable columns.</b> Columns 67 to 79, <span class=\"gene\">yibF</span> to <span class=\"gene\">ulaD</span>, are carried by 7 to 62%. The last column is <span class=\"gene\">ulaD</span>, in 18% of chromosomes.</p><p class=\"note\"><b>Across the window it names 83% of genes</b>, 0.52 bits on average, perplexity 2.2 families. Inside a region of plasticity it hesitates more than outside, 1.22 against 0.20 bits, and names fewer genes, 61% against 93%.</p>",
+ "window": "G8B09_RS13610",
+ "label": "yicS",
+ "facts_hash": "a4dd71e75b8c",
+ "prompt_version": "r9",
+ "skill_sha": "4a7fdc464826",
+ "facts": "[W] WINDOW: the yicS window: 80 columns, column 0 is the anchor yicS, then the next 79 genes in dnaA's direction; about genes 41 to 120 from dnaA. Read in 533 of the 540 complete chromosomes (those carrying yicS once); 42,639 calls of the model, each on a chromosome's real proteins.\n[T] TERMS: share: chromosomes carrying the family in the window. named: calls where the model names the gene at that column from the genes before it. bits: the model's entropy there, its hesitation (the more bits, the more hesitant). perplexity (whole window only): how surprised the model is by the real gene. RGP N%: N% of the calls there are on a gene inside a region of genomic plasticity (panRGP). P/S/C: persistent, shell, cloud (PPanGGOLiN). Spine column: one family in 90% of chromosomes or more; fork: a second family in 10% or more; variable: neither. A persistent family is in nearly every chromosome: in a fork or variable column its lower share means it lies beyond the window's end in the other chromosomes (genes inserted upstream push it out); never call it rare or say they lack it.\n[G] WHOLE WINDOW: the model names 83% of genes, 0.52 bits on average (perplexity 2.2 families). By partition of the gene: persistent 27,281 calls, named 96%, 0.15 bits; shell 14,111 calls, named 64%, 0.98 bits; cloud 1,247 calls, named 24%, 3.43 bits. Outside an RGP 29,182 calls, named 93%, 0.20 bits; inside an RGP 13,457 calls, named 61%, 1.22 bits. Columns: 29 spine, 16 fork, 35 variable.\n[C] COMPARISON (computed, over calls): inside an RGP the model hesitates more than outside (1.22 against 0.20 bits) and names fewer genes (61% against 93%); 13,457 of the 42,639 calls are inside an RGP.\n[H] HESITATION: the most hesitant columns of the window: column 4 (yicJ, a spine column) 2.66 bits, RGP 40%; column 5 (yicI, a spine column) 1.15 bits, RGP 41%; column 3 (setC, a fork column) 1.07 bits, RGP 40%. The 12 columns (columns 32-37 and 74-79) where most calls are in an RGP: named 75%, 0.43 bits on average; the other 68 columns: named 85%, 0.53 bits.\n[M] BY KIND OF COLUMN (computed, over the columns with calls): the 29 spine columns: named 85%, 0.71 bits on average; the 51 fork and variable columns: named 82%, 0.41 bits. The model names more genes on the spine columns than on the fork and variable ones, and hesitates more there. Its most hesitant column, 4, is a spine column: the model can hesitate most inside a spine.\n[P] PLASTICITY (the stretches below, by their mean over columns): the stretch most in regions of plasticity: columns 4-15 (spine), RGP 29-50%; the least: columns 0-2 (spine), RGP 0.2-2%.\n[L] STRETCHES, in window order (each spine; between spines, the fork and variable columns; the family listed is the most carried at its column; share of chromosomes; RGP: range over the columns):\n[S1] 1. columns 0-2 (3 columns) spine: share 97-100%, all P, RGP 0.2-2%. Genes: yicS \"protein YicS\", nlpA, yicL \"carboxylate/amino acid/amine transp\u2026\". named 95-99% (mean 97%), fewest named and most hesitant at column 1 (nlpA), named 95%, 0.21 bits.\n[S2] 2. column 3 fork: share 73%, all S, RGP 40%. named 83%; 1.07 bits.\n[S2.c3]    3: setC S 73%; also \"IS1 family transposase\" S 32%, \"pas-pac-pac sensing his kinase\" S 11%\n[S3] 3. columns 4-15 (12 columns) spine: share 90-98%, all P, RGP 29-50%. Genes: yicJ \"glycoside-pentoside-hexuronide fami\u2026\", yicI \"alpha-xylosidase\", yicH \"AsmA family protein\", xanP, gltS, recG, trmH, spoT, rpoZ, gmk, ligB, yicG \"trimeric intracellular cation chann\u2026\". named 57-91% (mean 83%), fewest named and most hesitant at column 4 (yicJ), named 57%, 2.66 bits.\n[S4] 4. column 16 variable: share 77%, all S, RGP 31%. named 83%; 0.80 bits.\n[S4.c16]    16: dinD S 77%\n[S5] 5. columns 17-28 (12 columns) spine: share 93-95%, all P, RGP 25-30%. Genes: yicC \"YicC/YloC family endoribonuclease\", rph, pyrE, slmA, dut, coaBC, radC, rpmB \"50S ribosomal protein L28\", rpmG, mutM, coaD, waaA. named 78-87% (mean 84%), fewest named and most hesitant at column 17 (yicC), named 78%, 0.86 bits.\n[S6] 6. column 29 fork: share 70%, all S, RGP 25%. named 77%; 0.56 bits.\n[S6.c29]    29: rfaQ S 70%; also rfaQ (another family) S 21%\n[S7] 7. columns 30-31 (2 columns) spine: share 91%, all P, RGP 26-33%. Genes: waaG, rfaP. named 83-85% (mean 84%), fewest named and most hesitant at column 31 (rfaP), named 83%, 0.57 bits.\n[S8] 8. columns 32-79 (48 columns) fork and variable columns: share 7-87%, 29 P, 19 S, RGP 11-63%. named 62-92% (mean 82%), fewest named at column 36, named 62%; most hesitant at column 65, 0.60 bits.\n[S8.b32]    a second branch of 5 families through columns 32-36, a parallel path that some chromosomes carry instead of the most carried families there, each following the previous one in most of its chromosomes, share 16%, all S; the branch's families: waaO, rfaY, \"UDP-glucose--(galactosyl) LPS alpha1,2-\u2026\", \"lipopolysaccharide N-acetylglucosaminyl\u2026\", rfaL\n[S8.c32]    32 fork: waaO S 50%; also \"glycosyltransferase family 8 protein\" S 14%\n[S8.c33]    33 fork: \"glycosyltransferase family 8 protein\" S 49%; also waaB S 23%\n[S8.c34]    34 fork: rfaY S 49%; also waaO S 23%\n[S8.c35]    35 fork: \"UDP-galactose--(galactosyl) LPS alpha1,\u2026\" S 48%; also rfaJ S 22%\n[S8.c36]    36 fork: \"glycosyltransferase family 2 protein\" S 32%; also rfaY S 22%, \"tyrosine-type recombinase/integrase\" S 10%\n[S8.c37]    37 fork: \"O-antigen ligase family protein\" S 32%; also waaZ S 22%\n[S8.c38]    38 variable: \"lipopolysaccharide N-acetylglucosaminyl\u2026\" S 14%\n[S8.c39]    39 variable: rfaL S 13%\n[S8.c40]    40 variable: rfaC P 87%\n[S8.c41]    41 fork: rfaF P 86% \"ADP-heptose--LPS heptosyltransferas\u2026\"; also yibB S 26%\n[S8.c42]    42 variable: rfaD P 86%\n[S8.c43]    43 variable: kbl P 85%\n[S8.c44]    44 fork: tdh P 85%; also \"glycosyltransferase family 25 protein\" S 16%\n[S8.c45]    45 fork: waaH P 82%; also \"O-antigen ligase family protein\" S 16%\n[S8.c46-56]    columns 46-56 variable, share 80-85%, all P: yibQ \"divergent polysaccharide deacetylas\u2026\" 85%, envC 84%, gpmM 84%, yibN \"rhodanese-like domain-containing pr\u2026\" 84%, grxC 83%, secB 83%, gpsA 82%, cysE 82%, trmL 81%, lldD \"quinone-dependent L-lactate dehydro\u2026\" 80%, lldR 80%\n[S8.c57]    57 fork: lldP P 79%; also ehaG S 11%\n[S8.c58]    58 fork: yibL P 77% \"YibL family ribosome-associated pro\u2026\"; also \"DUF3251 domain-containing protein\" S 21%\n[S8.c59-63]    columns 59-63 variable, share 69-75%, all P: yibT \"protein YibT\" 75%, mtlR \"mannitol operon repressor MtlR\" 73%, mtlD 72%, mtlA 70%, yibI \"DUF3302 domain-containing protein\" 69%\n[S8.c64]    64 fork: yibH P 68% \"HlyD family secretion protein\"; also \"YadA-like family protein\" S 13%\n[S8.c65]    65 fork: yibG S 39% \"sel1 repeat family protein\"; also \"DUF3251 domain-containing protein\" S 18%\n[S8.c66]    66 fork: rhsA S 38%; also yibA S 35%\n[S8.c67-79]    columns 67-79 variable, share 7-62%, 9 S, 4 P: yibF \"glutathione S-transferase\" 62%, selA 60%, selB \"selenocysteine-specific translation\u2026\" 58%, yiaY \"L-threonine dehydrogenase\" 51%, aldB 51%, \"hypothetical protein\" 30%, \"Imm57 family immunity protein\" 7%, yiaV \"HlyD family secretion protein\" 36%, yiaU \"LysR family transcriptional regulat\u2026\" 34%, yiaT \"MipA/OmpV family protein\" 33%, araD 32%, sgbU 25%, ulaD 18%\n[Q] GROUPS whose products agree, the places for a [[ ]] (name that group's columns exactly, cite its line and [Q], make one phrase from its words): columns 28-32 (waaA to waaO) [S5, S6, S7, S8]: lipopolysaccharide, core; the branch [S8.b32] through columns 32-36: lipopolysaccharide; columns 39-42 (rfaL to rfaD) [S8]: heptose, heptosyltransferase; columns 55-57 (lldD to lldP) [S8]: l-lactate, lactate; columns 60-62 (mtlR to mtlA) [S8]: mannitol.\n[X] PRODUCTS of named genes (for [[ ]]): rpmG 50S ribosomal protein L33; rfaL O-antigen ligase RfaL; rfaC lipopolysaccharide heptosyltransferase\u2026; rfaF ADP-heptose--LPS heptosyltransferase Rf\u2026; rfaD ADP-glyceromanno-heptose 6-epimerase; lldP L-lactate permease; mtlD mannitol-1-phosphate 5-dehydrogenase; mtlA PTS mannitol transporter subunit IICBA; selA L-seryl-tRNA(Sec) selenium transferase; cbeA type IV toxin-antitoxin system YeeU fam\u2026; rfaQ lipopolysaccharide core heptosyltransfe\u2026; waaO lipopolysaccharide 3-alpha-galactosyltr\u2026; waaB lipopolysaccharide 1,6-galactosyltransf\u2026; rfaY lipopolysaccharide core heptose(II) kin\u2026; rfaJ lipopolysaccharide 1,2-glucosyltransfer\u2026; waaZ 3-deoxy-D-manno-oct-2-ulosonate III tra\u2026; ehaG trimeric autotransporter adhesin EhaG\n[E] END: the window ends at column 79, ulaD, in 18% of chromosomes."
+}