Varia: a tool for prediction, analysis and visualisation of variable genes

Area of Science:

• Genomics
• Parasitology
• Bioinformatics

Background:

• Parasites utilize polymorphic gene families for immune evasion and host interaction.
• Plasmodium falciparum var genes encode the PfEMP1 protein, crucial for virulence and host receptor binding via DBL and CIDR domains.
• Assessing var gene diversity and expression is challenging but vital for understanding host-parasite interactions.

Purpose of the Study:

• To develop a computational tool, Varia, for predicting near full-length var gene sequences and domain compositions from short sequence tags.
• To enable rapid profiling of var gene expression in patient samples using DBLα expression sequence tags (EST).

Main Methods:

• Varia employs two pipelines: Varia_VIP for predicting full-length gene sequences and domain structures, and Varia_GEM for profiling var gene expression from DBLα ESTs.
• Varia_VIP analyzes partial sequences against database genes sharing short sequence tags.
• Varia_GEM computes transcript profiles stratified by PfEMP1 domain types.

Main Results:

• Varia_VIP successfully identified full-length query gene sequences for a significant portion of queries, with 92% of tags having highly identical database hits.
• Varia_GEM achieved high accuracy in predicting N-terminal domains (> 90%) and C-terminal domains (> 70%), enabling prediction for over 80% of queries.
• Prediction accuracy for C-terminal domains decreased with distance from the DBLα domain.

Conclusions:

• Computational prediction of var gene sequences and domain composition from short tags is feasible.
• The Varia tool facilitates experimental validation of PfEMP1 sequences and high-throughput analysis of var gene expression in clinical samples.