Workflow Intent
Perform transcriptomic profiling and differential gene expression analysis of human cell lines (A549 wild-type vs. A549 PTPN13 knockout) infected with Ebola virus versus mock infection across two post-infection time points (24 and 48 HPI). The primary objective is to assess the transcriptional host response, the effect of PTPN13 loss on viral infection kinetics and immune signalling, and time-dependent expression dynamics.
Methods / Algorithms
- Transcript Quantification: Pseudoalignment and transcript abundance estimation of paired-end RNA-seq reads against human reference transcriptome targets using Salmon.
- Gene-Level Summarization / Aggregation: Mapping transcript-level estimates to gene-level counts and abundances for downstream statistical modeling.
- Differential Expression Analysis: Multi-factor negative binomial generalized linear modeling (DESeq2) testing contrasts across Genotype (WT vs. PTPN13-KO), Infection Condition (Mock vs. Ebola virus), and Time Point (24 HPI vs. 48 HPI).
- Salmon: Paired-end read pseudoalignment and quantification (
salmon quant).
- DESeq2: Differential gene expression analysis and normalization from quantified counts.
- (Optional / Downstream helper): Gene-to-transcript mapping utility or tximport if converting transcript abundance to gene counts before DESeq2.
- RNA-seq Reads: Paired-end Illumina FASTQ reads formatted as a paired dataset collection or list of pairs (
list:paired):
- Cell lines: A549 WT and A549 PTPN13-KO
- Treatment: Mock vs. Ebola virus infection
- Time points: 24 HPI, 48 HPI
- Reference Transcriptome / Index: Human transcriptome sequences (FASTA) or precomputed Salmon index (GRCh38 / Ensembl / GENCODE).
- Sample Metadata / Design Matrix: Tabular factor annotations mapping sample identifiers to Genotype (WT / KO), Condition (Mock / Ebola), and Timepoint (24 / 48).
- Gene-Transcript Mapping Table: Two-column tabular mapping transcript IDs to gene symbols/IDs for DESeq2 gene aggregation.
Outputs
- Transcript & Gene Quantifications: Salmon quant directories / abundance tabular files (
quant.sf) per sample.
- Normalized Counts: Size-factor normalized expression matrix from DESeq2.
- Differential Expression Results: Tabular tables of log2 fold changes, p-values, and adjusted p-values (FDR) for key comparisons (e.g., Ebola vs Mock in WT, Ebola vs Mock in KO, Genotype effect upon infection).
- Diagnostic / QC Plots: PCA plot of samples, dispersion estimates, and MA plots generated by DESeq2.
Parameters
- Paired-end library orientation and automated strandedness detection (
-l A / automated).
- Default DESeq2 Wald test or LRT (likelihood ratio test) for multi-factor time series designs.
-
| Significance thresholds typically FDR < 0.05 and |
log2FC |
> 1. |
Data Availability
- User has experimental data; minimal synthetic/subsampled test fixtures (small FASTQ pairs + mini transcriptome index + sample sheet) will be resolved for automated workflow validation and testing.
Constraints
- Sequencing read format is paired-end Illumina.
- Organism focus: Human host transcriptome (A549 lung carcinoma line).
- Non-relevant noise terms (“PPR”, “DC”) mentioned during initial prompt were clarified as conversational noise and excluded.
Confidence and Open Questions
- Confidence: High confidence on paired-end read structure, Salmon quantification, human host reference, and DESeq2 differential calling across the 2x2x2 experimental matrix.
- Open Questions:
- Transcriptome vs Dual-Index: Does the Salmon index contain only the human transcriptome, or is it a combined host + Ebola virus transcriptome to concurrently quantify viral RNA load?
- Design Formula: Is the primary contrast modeled with interaction terms (e.g.
~ Genotype + Time + Condition + Genotype:Condition) or as grouped factor contrasts?
- Gene Mapping: Will the workflow receive a separate tx2gene tabular file, or use built-in tool features / annotations?