Predicting v(d)j recombination using conditional random fields

Authors:
Raunaq Malhotra;Shruthi Prabhakara;Raj Acharya
Affiliations:
Department of Computer Science Engineering, Pennsylvania State University, University Park, PA;Department of Computer Science Engineering, Pennsylvania State University, University Park, PA;Department of Computer Science Engineering, Pennsylvania State University, University Park, PA
Venue:
PRIB'12 Proceedings of the 7th IAPR international conference on Pattern Recognition in Bioinformatics
Year:
2012

Citing 9
Cited 0

A limited memory algorithm for bound constrained optimization

SIAM Journal on Scientific Computing
Conditional Random Fields: Probabilistic Models for Segmenting and Labeling Sequence Data

ICML '01 Proceedings of the Eighteenth International Conference on Machine Learning
Ultraconservative online algorithms for multiclass problems

The Journal of Machine Learning Research
SoDA: implementation of a 3D alignment algorithm for inference of antigen receptor recombinations

Bioinformatics
Protein--protein interaction site prediction based on conditional random fields

Bioinformatics
iHMMune-align

Bioinformatics
SoDA2

Bioinformatics
Search and clustering orders of magnitude faster than BLAST

Bioinformatics
Benchmarking the performance of human antibody gene alignment utilities using a 454 sequence dataset

Bioinformatics

Quantified Score

Hi-index	0.00

Visualization

Abstract

V(D)J gene segments undergo combinatorial recombination in the T-cells and B-cells to provide humans and other vertebrates with a large number of antibodies required for immunity. Each such recombination further undergoes mutations in their DNA sequences so that they can recognize diverse antigens. Predicting the combination of gene segments which formed a particular antibody is an essential task for studying disease propagation and analysis. We propose a model based on conditional random fields (CRFs) for predicting the boundary positions between V-D-J gene segments. We train the CRFs by generating synthetic gene recombinations using all of the alleles of the V, D and J gene segments. The alleles corresponding to a read can be determined by mapping the segmented reads to the DNA sequences of the gene segments using softwares like BLAST and usearch. We test our method on simulated dataset as well as real data of Stanford_S22 individual.