Performance of random forests and logic regression methods using mini-exome sequence data.

Machine learning approaches are an attractive option for analyzing large-scale data to detect genetic variants that contribute to variation of a quantitative trait, without requiring specific distributional assumptions. We evaluate two machine learning methods, random forests and logic regression, and compare them to standard simple univariate linear regression, using the Genetic Analysis Workshop 17 mini-exome data. We also apply these methods after collapsing multiple rare variants within genes and within gene pathways. Linear regression and the random forest method performed better when rare variants were collapsed based on genes or gene pathways than when each variant was analyzed separately. Logic regression performed better when rare variants were collapsed based on genes rather than on pathways.

Comparison of results from tests of association in unrelated individuals with uncollapsed and collapsed sequence variants using tiled regression.

Tiled regression is an approach designed to determine the set of independent genetic variants that contribute to the variation of a quantitative trait in the presence of many highly correlated variants. In this study, we evaluate the statistical properties of the tiled regression method using the Genetic Analysis Workshop 17 data in unrelated individuals for traits Q1, Q2, and Q4. To increase the power to detect rare variants, we use two methods to collapse rare variants and compare the results with those from the uncollapsed data. In addition, we compare the tiled regression method to traditional tests of association with and without collapsed rare variants. The results show that collapsing rare variants generally improves the power to detect associations regardless of method, although only variants with the largest allelic effects could be detected. However, for traditional simple linear regression, the average estimated type I error is dependent on the trait and varies by about three orders of magnitude. The estimated type I error rate is stable for tiled regression across traits.

A twisted hand: bHLH protein phosphorylation and dimerization regulate limb development.

Saethre-Chotzen syndrome (SCS), a human autosomal dominant condition with limb defects and craniosynostosis, is caused by haploinsufficiency of TWIST1, a basic helix-loop-helix (bHLH) transcription factor. Until recently, the molecular pathogenesis of the limb defects in SCS has not been well understood. Now, Firulli et al.1 show in mouse and chick that ectopic expression of a related bHLH protein, Hand2, results in phenocopies of the limb defects caused by Twist1 loss-of-function mutations. These two proteins interact in a dosage-dependent antagonistic manner, and both can be regulated through phosphorylation at conserved helix I amino acid residues. These findings provide an important link between the misregulation of Twist1 dimerization and the limb phenotypes observed in SCS.

Abnormalities in cartilage and bone development in the Apert syndrome FGFR2(+/S252W) mouse.

Apert syndrome is an autosomal dominant disorder characterized by malformations of the skull, limbs and viscera. Two-thirds of affected individuals have a S252W mutation in fibroblast growth factor receptor 2 (FGFR2). To study the pathogenesis of this condition, we generated a knock-in mouse model with this mutation. The Fgfr2(+/S252W) mutant mice have abnormalities of the skeleton, as well as of other organs including the brain, thymus, lungs, heart and intestines. In the mutant neurocranium, we found a midline sutural defect and craniosynostosis with abnormal osteoblastic proliferation and differentiation. We noted ectopic cartilage at the midline sagittal suture, and cartilage abnormalities in the basicranium, nasal turbinates and trachea. In addition, from the mutant long bones, in vitro cell cultures grown in osteogenic medium revealed chondrocytes, which were absent in the controls. Our results suggest that altered cartilage and bone development play a significant role in the pathogenesis of the Apert syndrome phenotype.

Gene expression in pharyngeal arch 1 during human embryonic development.

Craniofacial abnormalities are one of the most common birth defects in humans, but little is known about the human genes that control these important developmental processes. To identify relevant genes, we analyzed transcription profiles of human pharyngeal arch 1 (PA1), a conserved embryonic structure that develops into the palate and jaw. Using microdissected, normal human craniofacial structures, we constructed 12 SAGE (serial analysis of gene expression) libraries and sequenced 606 532 tags. We also performed Affymetrix microarray analysis on 25 craniofacial targets. Our data revealed not only genes "enriched" or differentially expressed in PA1 during fourth and fifth week of human development, but also 6927 genes newly identified to be expressed in human PA1. Many of these genes are involved in biosynthetic processes and have binding function and catalytic activity. We compared expression profiles of human genes with those of mouse homologs to look for genes more specific to human craniofacial development and found 766 genes expressed in human PA1, but not in mouse PA1. We also identified 1408 genes that were expressed in mouse as well as human PA1 and could be useful in creating mouse models for human conditions. We confirmed conservation of some human PA1 expression patterns in mouse embryonic samples with whole mount in situ hybridization and real-time RT-PCR. This comprehensive approach to expression profiling gives insights into the early development of the craniofacial region and provides markers for developmental structures and candidate genes, including SET and CCT3, for diseases such as orofacial clefting and micrognathia.

Increased risk for developmental delay in Saethre-Chotzen syndrome is associated with TWIST deletions: an improved strategy for TWIST mutation screening.

The majority of patients with Saethre-Chotzen syndrome have mutations in the TWIST gene, which codes for a basic helix-loop-helix transcription factor. Of the genetic alterations identified in TWIST, nonsense mutations, frameshifts secondary to small deletions or insertions, and large deletions implicate haploinsufficiency as the pathogenic mechanism. We identified three novel intragenic mutations and six deletions in our patients by using a new strategy to screen for TWIST mutations. We used polymerase chain reaction (PCR) amplification with subsequent sequencing to identify point mutations and small insertions or deletions in the coding region, and real-time PCR-based gene dosage analysis to identify large deletions encompassing the gene, with confirmation by microsatellite and fluorescence in situ hybridization (FISH) analyses. The size of the deletions can also be analyzed by using the gene dosage assay with "PCR walking" across the critical region. In 55 patients with features of Saethre-Chotzen syndrome, 11% were detected to have deletions by real-time gene dosage analysis. Two patients had a translocation or inversion at least 260 kb 3' of the gene, suggesting they had position-effect mutations. Of the 37 patients with classic features of Saethre-Chotzen syndrome, the overall detection rate for TWIST mutations was 68%. The risk for developmental delay in patients with deletions involving the TWIST gene is approximately 90% or eight times more common than in patients with intragenic mutations.