Measuring science: An exploration

This paper examines the available United States data on academic research and development (R&D) expenditures and the number of papers published and the number of citations to these papers as possible measures of “output” of this enterprise. We look at these numbers for science and engineering as a whole, for five selected major fields, and at the individual university field level. The published data in Science and Engineering Indicators imply sharply diminishing returns to academic R&D using published papers as an “output” measure. These data are quite problematic. Using a newer set of data on papers and citations, based on an “expanding” set of journals and the newly released Bureau of Economic Analysis R&D deflators, changes the picture drastically, eliminating the appearance of diminishing returns but raising the question of why the input prices of academic R&D are rising so much faster than either the gross domestic product deflator or the implicit R&D deflator in industry. A production function analysis of such data at the individual field level follows. It indicates significant diminishing returns to “own” R&D, with the R&D coefficients hovering around 0.5 for estimates with paper numbers as the dependent variable and around 0.6 if total citations are used as the dependent variable. When we substitute scientists and engineers in place of R&D as the right-hand side variables, the coefficient on papers rises from 0.5 to 0.8, and the coefficient on citations rises from 0.6 to 0.9, indicating systematic measurement problems with R&D as the sole input into the production of scientific output. But allowing for individual university field effects drives these numbers down significantly below unity. Because in the aggregate both paper numbers and citations are growing as fast or faster than R&D, this finding can be interpreted as leaving a major, yet unmeasured, role for the contribution of spillovers from other fields, other universities, and other countries.

Control of time-dependent biological processes by temporally patterned input

Temporal patterning of biological variables, in the form of oscillations and rhythms on many time scales, is ubiquitous. Altering the temporal pattern of an input variable greatly affects the output of many biological processes. We develop here a conceptual framework for a quantitative understanding of such pattern dependence, focusing particularly on nonlinear, saturable, time-dependent processes that abound in biophysics, biochemistry, and physiology. We show theoretically that pattern dependence is governed by the nonlinearity of the input–output transformation as well as its time constant. As a result, only patterns on certain time scales permit the expression of pattern dependence, and processes with different time constants can respond preferentially to different patterns. This has implications for temporal coding and decoding, and allows differential control of processes through pattern. We show how pattern dependence can be quantitatively predicted using only information from steady, unpatterned input. To apply our ideas, we analyze, in an experimental example, how muscle contraction depends on the pattern of motorneuron firing.

Sequence dependent hypermutation of the immunoglobulin heavy chain in cultured B cells

The variable (V) regions of immunoglobulin heavy and light chains undergo high rates of somatic mutation during the immune response. Although point mutations accumulate throughout the V regions and their immediate flanking sequences, analysis of large numbers of mutations that have arisen in vivo reveal that the triplet AGC appears to be most susceptible to mutation. We have stably transfected B cell lines with γ2a heavy chain constructs containing TAG nonsense codons in their V regions that are part of either a putative (T)AGC hot spot or a (T)AGA non-hot spot motif. Using an ELISA spot assay to detect revertants and fluctuation analysis to determine rates of mutation, the rate of reversion of the TAG nonsense codon has been determined for different motifs in different parts of the V region. In the NSO plasma cell line, the (T)AGC hot spot motif mutates at rates of ≈6 × 10−4/bp per generation and ≈3 × 10−5/bp per generation at residues 38 and 94 in the V region. At each of these locations, the (T)AGC hot spot motif is 20–30 times more likely to undergo mutation than the (T)AGA non-hot spot motif. Moreover, the AGA non-hot spot motif mutates at as high a rate as the hot spot motif when it is located adjacent to hot spot motifs, suggesting that more extended sequences influence susceptibility to mutation.

A serotonin transporter gene intron 2 polymorphic region, correlated with affective disorders, has allele-dependent differential enhancer-like properties in the mouse embryo

Polymorphic regions consisting of a variable number of tandem repeats within intron 2 of the gene coding for the serotonin transporter protein 5-HTT have been associated with susceptibility to affective disorders. We have cloned two of these intronic polymorphisms, Stin2.10 and Stin2.12, into an expression vector containing a heterologous minimal promoter and the bacterial LacZ reporter gene. These constructs were then used to produce transgenic mice. In embryonic day 10.5 embryos, both Stin2.10 and Stin2.12 produced consistent β-galactosidase expression in the embryonic midbrain, hindbrain, and spinal cord floor plate. However, we observed that the levels of β-galactosidase expression produced by both the Stin2.10 and Stin2.12 within the rostral hindbrain differed significantly at embryonic day 10.5. Our data suggest that these polymorphic variable number of tandem repeats regions act as transcriptional regulators and have allele-dependent differential enhancer-like properties within an area of the hindbrain where the 5-HTT gene is known to be transcribed at this stage of development.

Noise in neurons is message dependent

Neuronal responses are conspicuously variable. We focus on one particular aspect of that variability: the precision of action potential timing. We show that for common models of noisy spike generation, elementary considerations imply that such variability is a function of the input, and can be made arbitrarily large or small by a suitable choice of inputs. Our considerations are expected to extend to virtually any mechanism of spike generation, and we illustrate them with data from the visual pathway. Thus, a simplification usually made in the application of information theory to neural processing is violated: noise is not independent of the message. However, we also show the existence of error-correcting topologies, which can achieve better timing reliability than their components.

Analysis of variable (diversity) joining recombination in DNAdependent protein kinase (DNA-PK)-deficient mice reveals DNA-PK-independent pathways for both signal and coding joint formation

Previous studies have suggested that ionizing radiation causes irreparable DNA double-strand breaks in mice and cell lines harboring mutations in any of the three subunits of DNA-dependent protein kinase (DNA-PK) (the catalytic subunit, DNA-PKcs, or one of the DNA-binding subunits, Ku70 or Ku86). In actuality, these mutants vary in their ability to resolve double-strand breaks generated during variable (diversity) joining [V(D)J] recombination. Mutant cell lines and mice with targeted deletions in Ku70 or Ku86 are severely compromised in their ability to form coding and signal joints, the products of V(D)J recombination. It is noteworthy, however, that severe combined immunodeficient (SCID) mice, which bear a nonnull mutation in DNA-PKcs, are substantially less impaired in forming signal joints than coding joints. The current view holds that the defective protein encoded by the murine SCID allele retains enough residual function to support signal joint formation. An alternative hypothesis proposes that DNA-PKcs and Ku perform different roles in V(D)J recombination, with DNA-PKcs required only for coding joint formation. To resolve this issue, we examined V(D)J recombination in DNA-PKcs-deficient (SLIP) mice. We found that the effects of this mutation on coding and signal joint formation are identical to the effects of the SCID mutation. Signal joints are formed at levels 10-fold lower than in wild type, and one-half of these joints are aberrant. These data are incompatible with the notion that signal joint formation in SCID mice results from residual DNA-PKcs function, and suggest a third possibility: that DNA-PKcs normally plays an important but nonessential role in signal joint formation.

Transplantation of a 17-amino acid alpha-helical DNA-binding domain into an antibody molecule confers sequence-dependent DNA recognition.

Recombinant antibodies capable of sequence-specific interactions with nucleic acids represent a class of DNA- and RNA-binding proteins with potential for broad application in basic research and medicine. We describe the rational design of a DNA-binding antibody, Fab-Ebox, by replacing a variable segment of the immunoglobulin heavy chain with a 17-amino acid domain derived from TFEB, a class B basic helix-loop-helix protein. DNA-binding activity was studied by electrophoretic mobility-shift assays in which Fab-Ebox was shown to form a specific complex with DNA containing the TFEB recognition motif (CACGTG). Similarities were found in the abilities of TFEB and Fab-Ebox to discriminate between oligodeoxyribonucleotides containing altered recognition sequences. Comparable interference of binding by methylation of cytosine residues indicated that Fab-Ebox and TFEB both contact DNA through interactions along the major groove of double-stranded DNA. The results of this study indicate that DNA-binding antibodies of high specificity can be developed by using the modular nature of both immunoglobulins and transcription factors.