TACT

TACT annotates transcript sequences by integrating FASTY and BLASTX sequence similarity searches with InterProScan motif prediction to predict open reading frames (ORFs) and infer protein function for functional annotation of transcripts.

Key Features:

• Sequence similarity searches: Uses FASTY and BLASTX against UniProtKB (Swiss-Prot/TrEMBL) and RefSeq to detect homologous proteins and support ORF identification in mRNA sequences.
• Functional motif prediction: Employs InterProScan to identify conserved protein domains and functional motifs within predicted translations.
• Automated ORF-prediction pipeline: Automates prediction of plausible ORFs and generation of corresponding protein annotations.
• Database integration: Integrates annotations and data-mining with H-InvDB for comparison and consolidation of transcript information.
• Detection of structural features: Identifies SCOP domains and transmembrane regions within hypothetical proteins to inform structural and functional inference.
• Validation metrics: Includes benchmarking against human-curated annotations with an reported agreement rate of 83.9% on 19,574 non-redundant human transcripts.

Scientific Applications:

• Functional annotation of transcripts: Assigns putative protein functions to transcripts and hypothetical proteins by combining homology and motif evidence.
• ORF discovery: Predicts plausible open reading frames in mRNA sequences to prioritize coding candidates for downstream analysis.
• Structural-functional inference: Supports identification of SCOP domains and transmembrane proteins to enable structural predictions and functional hypothesis generation.
• Large-scale cDNA annotation: Applies to annotation of human full-length cDNA datasets such as those from the H-Inv project to produce high-throughput functional annotations.

Methodology:

mRNA sequences are subjected to FASTY and BLASTX searches against UniProtKB (Swiss-Prot/TrEMBL) and RefSeq, followed by InterProScan motif/domain prediction and automated ORF prediction, with annotations compared to H-InvDB human-curated entries for validation.