Upload README.md
Browse files
README.md
CHANGED
|
@@ -122,9 +122,9 @@ per-class F1 scores of 0.95 (low-risk), 0.74 (moderate), and 0.49
|
|
| 122 |
|
| 123 |
| Table | Description |
|
| 124 |
|-------|-------------|
|
| 125 |
-
| `oligos.csv` | Oligonucleotide identifiers, target genes, modality, 5'-to-3' sequence, sequence length, control assignment, source role |
|
| 126 |
| `chemical_modifications.csv` | Aggregate chemistry per oligo: sugar chemistry, backbone class, phosphorothioate fraction, modification pattern, GalNAc conjugation, purity, characterization-method metadata |
|
| 127 |
-
| `oligo_modification_positions.csv` |
|
| 128 |
| `biophysical_properties.csv` | Predicted Tm, ΔG, GC content, off-target hybridization burden, sequence-derived risk fields |
|
| 129 |
| `dose_exposure.csv` | Dose/concentration, exposure duration, treatment frequency, exposure normalization |
|
| 130 |
| `assays.csv` | Model system, cell model, species/human-proxy flag, organoid/MPS status, replicate count, assay type |
|
|
@@ -189,18 +189,18 @@ The full calibration specification is recorded in
|
|
| 189 |
non-trivial classification problem with both common and rare strata;
|
| 190 |
it is not a claim about wet-lab oligonucleotide toxicity prevalence.
|
| 191 |
- Of the 1,120 rows in `oligos.csv`, 1,089 represent oligos with
|
| 192 |
-
explicit sequences
|
| 193 |
-
|
| 194 |
-
|
| 195 |
-
|
| 196 |
-
|
| 197 |
-
|
| 198 |
-
|
| 199 |
-
`
|
|
|
|
|
|
|
| 200 |
`extracellular_vesicle_context_not_oligo`,
|
| 201 |
-
`clinical-risk-score-context`).
|
| 202 |
-
sequence-bearing oligos should filter on `sequence_length > 0` or on
|
| 203 |
-
the modality field.
|
| 204 |
- Original source PDFs are not redistributed; sources are identified
|
| 205 |
through title-level metadata only.
|
| 206 |
|
|
|
|
| 122 |
|
| 123 |
| Table | Description |
|
| 124 |
|-------|-------------|
|
| 125 |
+
| `oligos.csv` | Oligonucleotide identifiers, target genes, modality, 5'-to-3' sequence, sequence length, `is_administered_oligo` flag, control assignment, source role |
|
| 126 |
| `chemical_modifications.csv` | Aggregate chemistry per oligo: sugar chemistry, backbone class, phosphorothioate fraction, modification pattern, GalNAc conjugation, purity, characterization-method metadata |
|
| 127 |
+
| `oligo_modification_positions.csv` | Position-level chemical modifications: per-position base, sugar modification, backbone linkage, region (e.g., gapmer wing/gap, siRNA seed region), terminal conjugate, and provenance |
|
| 128 |
| `biophysical_properties.csv` | Predicted Tm, ΔG, GC content, off-target hybridization burden, sequence-derived risk fields |
|
| 129 |
| `dose_exposure.csv` | Dose/concentration, exposure duration, treatment frequency, exposure normalization |
|
| 130 |
| `assays.csv` | Model system, cell model, species/human-proxy flag, organoid/MPS status, replicate count, assay type |
|
|
|
|
| 189 |
non-trivial classification problem with both common and rare strata;
|
| 190 |
it is not a claim about wet-lab oligonucleotide toxicity prevalence.
|
| 191 |
- Of the 1,120 rows in `oligos.csv`, 1,089 represent oligos with
|
| 192 |
+
explicit sequences and 31 represent context records without an
|
| 193 |
+
administered oligo: vehicle/mock controls, clinical risk-score
|
| 194 |
+
contexts, extracellular-vesicle delivery contexts, small-molecule
|
| 195 |
+
rescue contexts, and non-administered injury-model contexts. Each
|
| 196 |
+
row carries an `is_administered_oligo` boolean column for
|
| 197 |
+
unambiguous filtering. Users restricting analysis to sequence-bearing
|
| 198 |
+
oligos should filter on `is_administered_oligo == True` (equivalent
|
| 199 |
+
to `sequence_length > 0`). Context records additionally carry an
|
| 200 |
+
`oligo_modality` value that explicitly identifies their non-oligo
|
| 201 |
+
nature (e.g., `vehicle_mock`, `context_not_administered_oligo`,
|
| 202 |
`extracellular_vesicle_context_not_oligo`,
|
| 203 |
+
`clinical-risk-score-context`).
|
|
|
|
|
|
|
| 204 |
- Original source PDFs are not redistributed; sources are identified
|
| 205 |
through title-level metadata only.
|
| 206 |
|