(draft) En analysis

.gitignore

@@ -8,3 +8,6 @@ outputdir
 inputdir/all_phenopackets.zip
 inputdir/phenopacket-store/
 .openai_cache.db
+supplemental_data/
+outputdir_all_2024_07_04/
+stagedb

data/all/prompts/en/PMID_10571775_KSN_II_1_en-prompt.txt

@@ -0,0 +1,18 @@
+I am running an experiment on a clinical case report to see how your diagnoses compare with those of human experts.
+I am going to give you part of a medical case. In this case, you are “Dr. GPT-4”, an AI language model who is providing
+a diagnosis. Here are some guidelines. First, there is a single definitive diagnosis, and it is a diagnosis that is known
+today to exist in humans. The diagnosis is almost always confirmed by some sort of genetic test, though in rare cases
+when such a test does not exist for a diagnosis the diagnosis can instead be made using validated clinical criteria or
+very rarely just confirmed by expert opinion. After you read the case, I want you to give a differential diagnosis with
+a list of candidate diagnoses ranked by probability starting with the most likely candidate. Each candidate should be
+specified with disease name. For instance, if the first candidate is Branchiooculofacial syndrome and the second is
+Cystic fibrosis, provide this:
+
+1. Branchiooculofacial syndrome
+2. Cystic fibrosis
+
+This list should provide as many diagnoses as you think are reasonable. You do not need to explain your reasoning,
+just list the diagnoses. Here is the case:
+
+The proband was a female. Disease onset was not specified.
+She presented with Growth delay, Hypokalemia, Hyperchloremic metabolic acidosis, Alkaline urine, Distal renal tubular acidosis, Elliptocytosis, Impaired urinary acidification, and Decreased serum bicarbonate concentration. However, the following features were excluded: Elevated circulating creatinine concentration.

data/all/prompts/en/PMID_10571775_YAT_II_1_en-prompt.txt

@@ -0,0 +1,18 @@
+I am running an experiment on a clinical case report to see how your diagnoses compare with those of human experts.
+I am going to give you part of a medical case. In this case, you are “Dr. GPT-4”, an AI language model who is providing
+a diagnosis. Here are some guidelines. First, there is a single definitive diagnosis, and it is a diagnosis that is known
+today to exist in humans. The diagnosis is almost always confirmed by some sort of genetic test, though in rare cases
+when such a test does not exist for a diagnosis the diagnosis can instead be made using validated clinical criteria or
+very rarely just confirmed by expert opinion. After you read the case, I want you to give a differential diagnosis with
+a list of candidate diagnoses ranked by probability starting with the most likely candidate. Each candidate should be
+specified with disease name. For instance, if the first candidate is Branchiooculofacial syndrome and the second is
+Cystic fibrosis, provide this:
+
+1. Branchiooculofacial syndrome
+2. Cystic fibrosis
+
+This list should provide as many diagnoses as you think are reasonable. You do not need to explain your reasoning,
+just list the diagnoses. Here is the case:
+
+The proband was a male. Disease onset was not specified.
+He presented with Growth delay, Hypokalemia, Hyperchloremic metabolic acidosis, Alkaline urine, Distal renal tubular acidosis, Elliptocytosis, Impaired urinary acidification, and Decreased serum bicarbonate concentration. However, the following features were excluded: Elevated circulating creatinine concentration.

data/all/prompts/en/PMID_10580070_A_III_11_en-prompt.txt

@@ -0,0 +1,18 @@
+I am running an experiment on a clinical case report to see how your diagnoses compare with those of human experts.
+I am going to give you part of a medical case. In this case, you are “Dr. GPT-4”, an AI language model who is providing
+a diagnosis. Here are some guidelines. First, there is a single definitive diagnosis, and it is a diagnosis that is known
+today to exist in humans. The diagnosis is almost always confirmed by some sort of genetic test, though in rare cases
+when such a test does not exist for a diagnosis the diagnosis can instead be made using validated clinical criteria or
+very rarely just confirmed by expert opinion. After you read the case, I want you to give a differential diagnosis with
+a list of candidate diagnoses ranked by probability starting with the most likely candidate. Each candidate should be
+specified with disease name. For instance, if the first candidate is Branchiooculofacial syndrome and the second is
+Cystic fibrosis, provide this:
+
+1. Branchiooculofacial syndrome
+2. Cystic fibrosis
+
+This list should provide as many diagnoses as you think are reasonable. You do not need to explain your reasoning,
+just list the diagnoses. Here is the case:
+
+The proband was a female. Disease onset occurred when the proband was 46-year, 0-month old.
+She presented with Atrial fibrillation, Elevated circulating creatine kinase concentration, and Second degree atrioventricular block. However, the following features were excluded: Dilated cardiomyopathy, Sudden cardiac death, Progeroid facial appearance, and Lipodystrophy.

data/all/prompts/en/PMID_10580070_A_III_13_en-prompt.txt

@@ -0,0 +1,18 @@
+I am running an experiment on a clinical case report to see how your diagnoses compare with those of human experts.
+I am going to give you part of a medical case. In this case, you are “Dr. GPT-4”, an AI language model who is providing
+a diagnosis. Here are some guidelines. First, there is a single definitive diagnosis, and it is a diagnosis that is known
+today to exist in humans. The diagnosis is almost always confirmed by some sort of genetic test, though in rare cases
+when such a test does not exist for a diagnosis the diagnosis can instead be made using validated clinical criteria or
+very rarely just confirmed by expert opinion. After you read the case, I want you to give a differential diagnosis with
+a list of candidate diagnoses ranked by probability starting with the most likely candidate. Each candidate should be
+specified with disease name. For instance, if the first candidate is Branchiooculofacial syndrome and the second is
+Cystic fibrosis, provide this:
+
+1. Branchiooculofacial syndrome
+2. Cystic fibrosis
+
+This list should provide as many diagnoses as you think are reasonable. You do not need to explain your reasoning,
+just list the diagnoses. Here is the case:
+
+The proband was a male. Disease onset occurred when the proband was 43-year, 0-month old.
+He presented with Atrial fibrillation, Sinus bradycardia, and First degree atrioventricular block. However, the following features were excluded: Dilated cardiomyopathy, Elevated circulating creatine kinase concentration, Sudden cardiac death, Progeroid facial appearance, and Lipodystrophy.

data/all/prompts/en/PMID_10580070_A_III_14_en-prompt.txt

@@ -0,0 +1,18 @@
+I am running an experiment on a clinical case report to see how your diagnoses compare with those of human experts.
+I am going to give you part of a medical case. In this case, you are “Dr. GPT-4”, an AI language model who is providing
+a diagnosis. Here are some guidelines. First, there is a single definitive diagnosis, and it is a diagnosis that is known
+today to exist in humans. The diagnosis is almost always confirmed by some sort of genetic test, though in rare cases
+when such a test does not exist for a diagnosis the diagnosis can instead be made using validated clinical criteria or
+very rarely just confirmed by expert opinion. After you read the case, I want you to give a differential diagnosis with
+a list of candidate diagnoses ranked by probability starting with the most likely candidate. Each candidate should be
+specified with disease name. For instance, if the first candidate is Branchiooculofacial syndrome and the second is
+Cystic fibrosis, provide this:
+
+1. Branchiooculofacial syndrome
+2. Cystic fibrosis
+
+This list should provide as many diagnoses as you think are reasonable. You do not need to explain your reasoning,
+just list the diagnoses. Here is the case:
+
+The proband was a female. Disease onset occurred when the proband was 40-year, 0-month old.
+She presented with First degree atrioventricular block. However, the following features were excluded: Atrial fibrillation, Dilated cardiomyopathy, Elevated circulating creatine kinase concentration, Progeroid facial appearance, and Lipodystrophy.

data/all/prompts/en/PMID_10580070_A_III_15_en-prompt.txt

@@ -0,0 +1,18 @@
+I am running an experiment on a clinical case report to see how your diagnoses compare with those of human experts.
+I am going to give you part of a medical case. In this case, you are “Dr. GPT-4”, an AI language model who is providing
+a diagnosis. Here are some guidelines. First, there is a single definitive diagnosis, and it is a diagnosis that is known
+today to exist in humans. The diagnosis is almost always confirmed by some sort of genetic test, though in rare cases
+when such a test does not exist for a diagnosis the diagnosis can instead be made using validated clinical criteria or
+very rarely just confirmed by expert opinion. After you read the case, I want you to give a differential diagnosis with
+a list of candidate diagnoses ranked by probability starting with the most likely candidate. Each candidate should be
+specified with disease name. For instance, if the first candidate is Branchiooculofacial syndrome and the second is
+Cystic fibrosis, provide this:
+
+1. Branchiooculofacial syndrome
+2. Cystic fibrosis
+
+This list should provide as many diagnoses as you think are reasonable. You do not need to explain your reasoning,
+just list the diagnoses. Here is the case:
+
+The proband was a female. Disease onset occurred when the proband was 39-year, 0-month old.
+She presented with Third degree atrioventricular block. However, the following features were excluded: Atrial fibrillation, Dilated cardiomyopathy, Sudden cardiac death, Progeroid facial appearance, and Lipodystrophy.

data/all/prompts/en/PMID_10580070_A_III_1_en-prompt.txt

@@ -0,0 +1,18 @@
+I am running an experiment on a clinical case report to see how your diagnoses compare with those of human experts.
+I am going to give you part of a medical case. In this case, you are “Dr. GPT-4”, an AI language model who is providing
+a diagnosis. Here are some guidelines. First, there is a single definitive diagnosis, and it is a diagnosis that is known
+today to exist in humans. The diagnosis is almost always confirmed by some sort of genetic test, though in rare cases
+when such a test does not exist for a diagnosis the diagnosis can instead be made using validated clinical criteria or
+very rarely just confirmed by expert opinion. After you read the case, I want you to give a differential diagnosis with
+a list of candidate diagnoses ranked by probability starting with the most likely candidate. Each candidate should be
+specified with disease name. For instance, if the first candidate is Branchiooculofacial syndrome and the second is
+Cystic fibrosis, provide this:
+
+1. Branchiooculofacial syndrome
+2. Cystic fibrosis
+
+This list should provide as many diagnoses as you think are reasonable. You do not need to explain your reasoning,
+just list the diagnoses. Here is the case:
+
+The proband was a male. Disease onset occurred when the proband was 51-year, 0-month old.
+He presented with Dilated cardiomyopathy and Third degree atrioventricular block. However, the following features were excluded: Atrial fibrillation, Elevated circulating creatine kinase concentration, Sudden cardiac death, Progeroid facial appearance, and Lipodystrophy.

data/all/prompts/en/PMID_10580070_A_III_5_en-prompt.txt

@@ -0,0 +1,18 @@
+I am running an experiment on a clinical case report to see how your diagnoses compare with those of human experts.
+I am going to give you part of a medical case. In this case, you are “Dr. GPT-4”, an AI language model who is providing
+a diagnosis. Here are some guidelines. First, there is a single definitive diagnosis, and it is a diagnosis that is known
+today to exist in humans. The diagnosis is almost always confirmed by some sort of genetic test, though in rare cases
+when such a test does not exist for a diagnosis the diagnosis can instead be made using validated clinical criteria or
+very rarely just confirmed by expert opinion. After you read the case, I want you to give a differential diagnosis with
+a list of candidate diagnoses ranked by probability starting with the most likely candidate. Each candidate should be
+specified with disease name. For instance, if the first candidate is Branchiooculofacial syndrome and the second is
+Cystic fibrosis, provide this:
+
+1. Branchiooculofacial syndrome
+2. Cystic fibrosis
+
+This list should provide as many diagnoses as you think are reasonable. You do not need to explain your reasoning,
+just list the diagnoses. Here is the case:
+
+The proband was a female. Disease onset occurred when the proband was 39-year, 0-month old.
+She was found not to have the following features: Atrial fibrillation, Dilated cardiomyopathy, Elevated circulating creatine kinase concentration, Sudden cardiac death, Progeroid facial appearance, and Lipodystrophy.

data/all/prompts/en/PMID_10580070_A_III_8__en-prompt.txt

@@ -0,0 +1,18 @@
+I am running an experiment on a clinical case report to see how your diagnoses compare with those of human experts.
+I am going to give you part of a medical case. In this case, you are “Dr. GPT-4”, an AI language model who is providing
+a diagnosis. Here are some guidelines. First, there is a single definitive diagnosis, and it is a diagnosis that is known
+today to exist in humans. The diagnosis is almost always confirmed by some sort of genetic test, though in rare cases
+when such a test does not exist for a diagnosis the diagnosis can instead be made using validated clinical criteria or
+very rarely just confirmed by expert opinion. After you read the case, I want you to give a differential diagnosis with
+a list of candidate diagnoses ranked by probability starting with the most likely candidate. Each candidate should be
+specified with disease name. For instance, if the first candidate is Branchiooculofacial syndrome and the second is
+Cystic fibrosis, provide this:
+
+1. Branchiooculofacial syndrome
+2. Cystic fibrosis
+
+This list should provide as many diagnoses as you think are reasonable. You do not need to explain your reasoning,
+just list the diagnoses. Here is the case:
+
+The proband was a female. Disease onset occurred when the proband was 42-year, 0-month old.
+She presented with Atrial fibrillation, Dilated cardiomyopathy, Thromboembolism, Sinus bradycardia, and Abnormal atrioventricular conduction. However, the following features were excluded: Elevated circulating creatine kinase concentration, Sudden cardiac death, Progeroid facial appearance, and Lipodystrophy.

data/all/prompts/en/PMID_10580070_A_III_9_en-prompt.txt

@@ -0,0 +1,18 @@
+I am running an experiment on a clinical case report to see how your diagnoses compare with those of human experts.
+I am going to give you part of a medical case. In this case, you are “Dr. GPT-4”, an AI language model who is providing
+a diagnosis. Here are some guidelines. First, there is a single definitive diagnosis, and it is a diagnosis that is known
+today to exist in humans. The diagnosis is almost always confirmed by some sort of genetic test, though in rare cases
+when such a test does not exist for a diagnosis the diagnosis can instead be made using validated clinical criteria or
+very rarely just confirmed by expert opinion. After you read the case, I want you to give a differential diagnosis with
+a list of candidate diagnoses ranked by probability starting with the most likely candidate. Each candidate should be
+specified with disease name. For instance, if the first candidate is Branchiooculofacial syndrome and the second is
+Cystic fibrosis, provide this:
+
+1. Branchiooculofacial syndrome
+2. Cystic fibrosis
+
+This list should provide as many diagnoses as you think are reasonable. You do not need to explain your reasoning,
+just list the diagnoses. Here is the case:
+
+The proband was a male. Disease onset occurred when the proband was 40-year, 0-month old.
+He presented with Dilated cardiomyopathy, Congestive heart failure, and Third degree atrioventricular block. However, the following features were excluded: Atrial fibrillation, Progeroid facial appearance, and Lipodystrophy.

data/all/prompts/en/PMID_10580070_B_III_11_en-prompt.txt

@@ -0,0 +1,18 @@
+I am running an experiment on a clinical case report to see how your diagnoses compare with those of human experts.
+I am going to give you part of a medical case. In this case, you are “Dr. GPT-4”, an AI language model who is providing
+a diagnosis. Here are some guidelines. First, there is a single definitive diagnosis, and it is a diagnosis that is known
+today to exist in humans. The diagnosis is almost always confirmed by some sort of genetic test, though in rare cases
+when such a test does not exist for a diagnosis the diagnosis can instead be made using validated clinical criteria or
+very rarely just confirmed by expert opinion. After you read the case, I want you to give a differential diagnosis with
+a list of candidate diagnoses ranked by probability starting with the most likely candidate. Each candidate should be
+specified with disease name. For instance, if the first candidate is Branchiooculofacial syndrome and the second is
+Cystic fibrosis, provide this:
+
+1. Branchiooculofacial syndrome
+2. Cystic fibrosis
+
+This list should provide as many diagnoses as you think are reasonable. You do not need to explain your reasoning,
+just list the diagnoses. Here is the case:
+
+The proband was a female. Disease onset occurred when the proband was 53-year, 0-month old.
+She presented with Atrial fibrillation, Dilated cardiomyopathy, Stroke, and Abnormal atrioventricular conduction. However, the following features were excluded: Elevated circulating creatine kinase concentration, Sudden cardiac death, Progeroid facial appearance, and Lipodystrophy.

data/all/prompts/en/PMID_10580070_B_III_13_en-prompt.txt

@@ -0,0 +1,18 @@
+I am running an experiment on a clinical case report to see how your diagnoses compare with those of human experts.
+I am going to give you part of a medical case. In this case, you are “Dr. GPT-4”, an AI language model who is providing
+a diagnosis. Here are some guidelines. First, there is a single definitive diagnosis, and it is a diagnosis that is known
+today to exist in humans. The diagnosis is almost always confirmed by some sort of genetic test, though in rare cases
+when such a test does not exist for a diagnosis the diagnosis can instead be made using validated clinical criteria or
+very rarely just confirmed by expert opinion. After you read the case, I want you to give a differential diagnosis with
+a list of candidate diagnoses ranked by probability starting with the most likely candidate. Each candidate should be
+specified with disease name. For instance, if the first candidate is Branchiooculofacial syndrome and the second is
+Cystic fibrosis, provide this:
+
+1. Branchiooculofacial syndrome
+2. Cystic fibrosis
+
+This list should provide as many diagnoses as you think are reasonable. You do not need to explain your reasoning,
+just list the diagnoses. Here is the case:
+
+The proband was a female. Disease onset occurred when the proband was 39-year, 0-month old.
+She presented with Atrial fibrillation, Dilated cardiomyopathy, Sudden cardiac death, Congestive heart failure, and Abnormal atrioventricular conduction. However, the following features were excluded: Progeroid facial appearance and Lipodystrophy.