> ## Documentation Index > Fetch the complete documentation index at: https://instance.bio/docs/llms.txt > Use this file to discover all available pages before exploring further. # binding_umi_counts Contains molecule-level sequencing read counts for candidates and observed sequence variants across samples that underwent antigen selection (binding samples). Before counting reads, alignments are filtered to include only reads that match the candidate sequence with zero errors. The number of reads observed per Unique Molecular Identifier (UMI) helps determine sequencing saturation for the sample. When each UMI is observed across many reads (for example, dozens or hundreds of times), sequencing is near saturation and additional sequencing depth is unlikely to capture new molecules. Conversely, when read counts per UMI are low, deeper sequencing may capture additional UMIs, allowing more candidate designs and rare sequence variants to be recovered. **Grain:** one row per `(binding_sample_id, variant_id, umi)`. | Column | Type | Description | | :- | :- | :- | | `binding_sample_id` | `string` | Identifies the binding sample this measurement is based on, matching *sample\_id* in `dim_samples`. | | `candidate_id` | `string` | Identifies this candidate across the dataset. Two candidates with identical sequences share the same `candidate_id`. | | `variant_id` | `string` | Identifies the specific sequence variant observed for this candidate. Equals `candidate_id` when the observed sequence matches the candidate sequence exactly. | | `umi` | `string` | Unique Molecular Identifier (UMI) sequence extracted from the read, identifying the single original physical molecule before PCR amplification. | | `read_count` | `integer` | Total number of sequencing reads that carried this `UMI`, reflecting the sequencing depth of this single physical molecule. | This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.