Read count vs confidence

Semi-quantitative interpretation of Nanopore detection signals

Note

Private This page documents internal logic for interpreting read count and confidence levels in detection workflows.

Overview

Nanopore sequencing 在 pathogen detection 中：

read count ≠ absolute quantification
但也 ≠ 沒有意義

👉 它是一種 semi-quantitative signal

Core concept

read count 反映：

  • abundance（部分）
  • sequencing depth
  • library bias
  • stochastic variation

因此：

read count 必須與其他指標一起解讀

Signal interpretation layers

Layer 1：Presence signal

  • 是否存在 reads？

👉 只能回答：

“可能存在”

Layer 2：Strength

  • read count
  • relative abundance

👉 提供：

signal 強度

Layer 3：Support

  • alignment quality
  • coverage

👉 提供：

是否可信

Layer 4：Context

  • sample type
  • control
  • biology

👉 決定：

是否有意義

Conceptual model

flowchart TD
    A[Read count] --> B[Signal strength]
    B --> C[Alignment validation]
    C --> D[Context evaluation]
    D --> E[Confidence level]

Internal interpretation tiers（可自行定義）

Tier Description Confidence
Tier 0 No detection None
Tier 1 Very low reads Very low
Tier 2 Moderate reads Suggestive
Tier 3 High reads + good alignment Strong

👉 ⚠️ threshold 需依專案定義

Key influencing factors

  • sequencing depth
  • host DNA proportion
  • library efficiency
  • genome size（target organism）
  • GC bias

Practical considerations

Low read count

可能：

  • early infection
  • low abundance organism
  • contamination

👉 必須搭配：

  • alignment
  • control

High read count

通常：

  • 較可信 但仍需確認：

  • specificity

  • context

Clinical vs environmental

Clinical sample

  • 偏向高 specificity
  • 低 read count 需謹慎

Environmental sample

  • 高 diversity
  • 低 read count 常見

Key principle

Tip

Read count 是訊號強度，而不是證據本身。

Internal notes（可填）

  • read count threshold：
  • coverage cutoff：
  • control rule：
  • reporting criteria：