In these platforms, two molecules that significantly differ in their masses represent the bits 0 and 1, respectively, and are used as building blocks in the process of synthesizing user-defined information content. The synthetic polymers are read by tandem mass (MS/MS) spectrometers. A mass spectrometer breaks multiple copies of the polymer uniformly at random, thereby fragmenting the string into substrings. The masses of these substrings are reported as the output of the system.

We propose a new family of codes that allows for both unique string reconstruction and correction of multiple mass errors.

Noting that these polymers are stored and read together, we design codebooks with asymptotically optimal rates such that multiple strings can be reconstructed based on their joint MS/MS readout. [arXiv 2003.02121]

Group testing was recently shown to increase the storage density of topological DNA-based data storage. In such a system, nanoscopic holes are punched into the sugar-phosphate backbone of one strand of a double-stranded DNA molecule. A "hole" indicates the value 1 while the absence of a hole indicates the value 0. Multiple copies of the same DNA strands are punched to bear different patterns. These are mixed together and stored in a single microwell. The mixture with various punch patterns are then sequenced and read together.

However, multiple holes punched in close proximity results in poorer quality outputs leading to a constraint on the runlength of 0s between pairs of 1s. Thus, we aim to construct a binary code such that multiple strings can be sequenced and read simultaneously with precision.

We address this problem by introducing a new runlength limited group testing paradigm, a simplification of the actual mixture identification problem. [DOI 10.1109/ISIT44484.2020.9174502]

In this collaborative work, we developed a comprehensive guide that highlights 1) the issues specifically posed by the Sars-Cov-2 virus, and 2) features of the RT-PCR testing mechanism and provide innovative algorithms to tackle the problem of testing under limited resources. In particular, we focus on the semiquantitative nature of the RT-PCR tests: not only do these tests detect the presence of the virus in the sample, these tests also provide partial information about the viral loads present in the sample. [arXiv:2011.05223]

Profile Hidden Markov Models (HMMs) are graphical models that can be used to produce finite length sequences from a distribution, with multiple applications, including protein structure and function prediction, and multiple sequence alignment. Although statistical identifiability is a fundamental aspect of any statistical model, identifiability of profile HMMs was not investigated previously. In our work, we reported on preliminary results towards characterizing the statistical identifiability of profile HMMs in one of the standard forms used in bioinformatics. [DOI 10.1109/TCBB.2019.2933821]