Wednesday 24 June 2020

Reading List: Rats on the latest celebrity diet

Caloric Restriction Reprograms the Single-Cell Transcriptional Landscape of Rattus Norvegicus Aging
 
Calorie Restriction (CR) leads to slow down of ageing in mammals. The statement has been passed around and been a cause for scientific debates. What is missing to substantiate the former claim is a CR cell atlas across body tissues. The research observes (in mice) the systemic effects of aging and CR on different tissues evaluated in terms of cell type composition, tissue-specific molecular programs, regulatory transcription factors (TFs), and cell-cell communication networks. It was observed that fewer lipid droplets accumulated in livers of the experimental group (excess lipids cause atherosclerosis: high cholesterol levels). An accumulation of senescent cells was found in the control group (senescence is caused by gradual telomere attrition at the ends of DNA reducing reproducibility) as compared to the experimental group. The number of immune cells in nearly every tissue studied dramatically increased as control rats aged but was not affected by age in rats with restricted calories. Levels of the transcription factor Ybx1 were altered by the diet in 23 different cell types (out of 40 chosen). The scientists believe Ybx1 may be an age-related transcription factor and are planning more research into its effects.


Interfacing gene circuits with microelectronics through engineered population dynamics

In this paper, the authors propose an alternative to fluorescent reporters to analyse bacterial population behaviours. They interfaced synthetic biology with microelectronics through engineered population dynamics that regulate the accumulation of charged metabolites. During bacterial growth, charged ions are naturally released because of metabolic processes and the environment becomes more conductive, which decreases the impedance to electrical current. To control bacterial populations, they engineered a genetic circuit to express a bacterial killing gene that is activated upon the addition of an external stimulus. Therefore the bacterial population resembles a resistor, which is controlled by a genetic circuit. The plan is to implement this regulatory system in a microelectronic platform where several chambers may contain unique genetic circuits, connected via electrodes to an impedance output system.


Isothermal digital detection of microRNAs using background-free molecular circuit

Micro-RNAs have emerged as a class of potential biomarkers due to an increasing amount of research identify their disregulation in several diseases. But detection of very low concentration of mi-RNA with high accuracy poses a challenge for the existing techniques. One of them is background noise and nonspecific amplification of products which renders low-concentration detection of microRNA practically impossible. This article describes the utilization of already existing PEN-DNA toolbox coupled with a leak absorption mechanism to eliminate background noise. This method generates a fluorescence signal specific for the amplification of the signal strand which is produced upstream by the microRNA. The overall system can detect microRNA concentration as low as 1fM.


Information-theoretical bound of the irreversibility in thermal relaxation processes

Relaxation processes must produce entropy as they cannot be quasistatic or reversible. Here Shiraishi dervied a stronger than 2nd law bound on the entropy production in for a relaxation process. The bound on the entropy production is given by the Kullback divergence between the initial and current time distribution of the system. Hence if the initial and final distributions are very different then there MUST be a large entropy production.
Brings up an interesting way of describing permissable trajectories using information geometry (interpret KL div as an euclidean distance).


Cancer Diagnosis with DNA molecular computation

A simplified winner takes all DNA computing scheme using 4 miRNA input was used to diagnose lung cancer cells. The network was trained in silico and realised with DNA.


The Synthesis Success Calculator: Predicting the Rapid Synthesis of DNA Fragments with Machine Learning

The efficiency of DNA synthesis is sequence-dependent; however, the effects of each sequence property is not well understood. To address this problem, the authors design a random forest classifier to quantify the effects of 38 sequence properties in the synthesis of more than one thousand DNA sequences. The conclusion is that only 9 properties are relevant, the most important being the length of the longest repetitive sequence within the DNA fragment (26 nt is the limit), followed by the strand GC content (between 29-63%). A predictive tool is available online to estimate the success of your DNA synthesis orders and suggest changes for synonym codons: https://salislab.net/software/


A Dynamical Biomolecular Neural Network 

Artificial neural networks (ANN) are amongst the most used computation models in machine learning and they have been proved to be extremely powerful for classification tasks in silico. Some DNA strand displacement implementations of these systems have been developed in the past, but this circuits were limited by its size, implementability and inability to be rebooted. In the present work, Ron Weiss and collaborators described a theoretical implementation of a perceptron (the functional unit of an ANN) in a biochemically feasable CRN. In the CRN two mutually sequestering chemical species with their production and degradation rates, encode the positive and negative weights of the perceptron.  Based in this design, the authors demonstrate that perceptrons can be extended into deeper networks and implement complex behaviours.


Small RNA driven feed-forward loop: Fine-tuning of protein synthesis through sRNA mediated cross-talk

Tej and Mukherji present an analysis of an sRNA-mediated feed-forward loop, in which a particular sRNA not only promotes translation of a protein, but also translation of a second protein (a sigma factor) that in turn promotes transcription of the mRNA of the first protein. They show that competition between the mRNA for the sRNA leads to a non-monotonic effect of sigma factor transcription rate on output protein levels, and that relative fluctuations are smallest at the point of maximal output protein levels.