National intercomparisons of I radioactivity measurements in nuclear medicine centres in India

National intercomparisons of activity measurements of ¹³¹I, a radioisotope widely used for diagnosis and therapy of thyroid related ailments, were initiated in 1979 as a quality assurance program, towards improving radiation safety procedures and related dosimetry in Nuclear Medicine Centres (NMCs) in India. Oral administration of a known quantity of radioiodine to patients requires accurate radioactivity measurements to be performed on a well-calibrated isotope calibrators. Under or over estimation of the activity due to a faulty or uncalibrated isotope calibrator could provide misleading results. Calibration of isotope calibrators and the traceablity of subsequent measurements to the national standards laboratory is one of the essential basic radiation safety requirement of the IAEA. In view of the stringent quality assurance requirements for activity measurements imposed by Atomic Energy Regulatory Board, a National Intercomparison Program was initiated and to date ten such intercomparison programs have been conducted by the Radiation Safety Systems Division, of the Bhabha Atomic Research Centre. This program has benefited the participants by making their measurements traceable to the National Primary Standards. Over the years there has been a marked increase in the number of NMCs participating in the intercomparison programs. As a result, the number of institution showing large deviation from the correct value has decreased considerably over the years. This program thus, has enabled participating NMCs to check their isotope calibrators so as to ensure proper delivery of radiation dose to the patients and hence to optimise patient exposure.     

Modeling of gene regulatory networks with hybrid differential evolution and particle swarm optimization.

In the last decade, recurrent neural networks (RNNs) have attracted more efforts in inferring genetic regulatory networks (GRNs), using time series gene expression data from microarray experiments. This is critically important for revealing fundamental cellular processes, investigating gene functions, and understanding their relations. However, RNNs are well known for training difficulty. Traditional gradient descent-based methods are easily stuck in local minima and the computation of the derivatives is also not always possible. Here, the performance of three evolutionary-swarm computation technology-based methods, known as differential evolution (DE), particle swarm optimization (PSO), and the hybrid of DE and PSO (DEPSO), in training RNNs is investigated. Furthermore, the gene networks are reconstructed via the identification of the gene interactions, which are explained through corresponding connection weight matrices. The experimental results on two data sets studied in this paper demonstrate that the DEPSO algorithm performs better in RNN training. Also, the RNN-based model can provide meaningful insight in capturing the nonlinear dynamics of genetic networks and revealing genetic regulatory interactions.