Home | ResearchCV | Publications | Personal |   

Computational annotation of GPI molecules


In my Ph.D. at UTEP I worked on the development of a new algorithm to interpret mass spectrometric (MS) data of the fragmentation of glycosylphosphatidylinositol (GPI)-anchored proteins (GPI-APs). I worked under the direction of Dr. Ming-Ying Leung from Bioinformatics and Computational Science department; and Dr. Igor C. Almeida from the department of Biological Sciences.

GPI-Anchoring is a universal post-translational modification (PTM) in eukaryotes and critical for embryonic development.  GPI-APs seem to be involved in numerous biological processes, including cell-cell interactions, signal transduction, endocytosis, complement regulation, and antigenic presentation. Nonetheless, surprisingly, the functional role of most GPI-APs and their GPI moieties in humans and many other mammals remains elusive, and only a handful of GPI anchors from human cells or tissues have been described thus far. The model organisms to study GPI-APs and protein-free GPIs are protozoan parasites, in particular of T. cruzi, the causative agent of Chagas disease, which affects millions of people mainly in Latin America. T. cruzi GPI-APs and protein-free GPIs (also known as GIPLs) are perhaps two of the most attractive molecular targets for vaccine and drug development in Chagas disease.

Below is an already annotated spectra.

Mass spectra

High Performance Computing

To reduce the waiting time for the GPI predictions, I implemented the algorithm in two separate parallel computing tools: HTCondor and MPI. HTCondor harnesses the idle CPU cycles in the grid of computers we have in our bioinformatics student computing lab on campus. Condor’s file transfer mechanisms distribute the workload to the available processors in the grid, which consists of 54 processors in 64-bit machines running CentOS Linux. The processor speeds range from 1 to 3.3 GHz Condor handles all the details of sending executable and data files to computing resources and retrieving the computation results. Also, Condor provides checkpointing: if the application is interrupted, checkpointing saves the computation’s state so it can be resumed later (instead of starting from scratch).

The implementation of MPI was done in The High Performance Computing Virtual Research Lab (HPCVRL).

Home | ResearchCV | Publications | Personal |