∪k-mer - knockout!

We hold a view that DNA sequences inherently follows strict rules of diffusion, spliced introns function on protein non-randomly and sequence order as well as length remain critical upstream of transcription. This formed the basis of our original data exploration in which we first proved the non-random relationships between intron's and the ultimate protein product of its gene.

In subsequence experiments we used our bioinformatic to identify a specific profile of HeLa cell sequences that reduced proliferation.  To discover these sequences, we computed more than 400,000 intron1 k-mers from 41 transcripts, identified and transfected 6 short sequences that significantly reduced the growth of these cervical cancer cells. From our RNA-seq analysis growth was reduced by shifting replicative senescence to apoptosis accompanied by mitcochondrial hyper-function.

For transcript (T), we looked at any sequence (S) as having cell-wide potential to define or be re-defined during or post transcription. However, as DNA of a gene, S must be considered to possess a unique potential that will effect its interactions. Science knows very little about the catalog of possible effects that can be attributed to S. To model its potential qualities we first look to its structural arrangements for any length (n>7).

In our examples below, for S1=TGTGGGCCCACA and S2=GTGGGCCCAGAC we computed every possible k(n>7) and count unions of k - ⋃k.

The Unions  of k for S1 and S2

Why is this important? Let's say you want to tie this computation to biological function, the base data would have to represent the myriad underlying biological possibilities. To do this you would need very low level analytical resolution, some of which is represented in the images above.

At Codondex we take a further step by associating every k with a hash signature (#s) of the transcript protein or mRNA. You will appreciate for T that #s is constant and associated with every k(n>7)  input to tuple k(n>7...S):#s. The combination of each intron 1 k-mer with its transcript protein signature becomes the input to a highly ordered vector used to compare order stability, in the vector at the next nucleotide of k-mer:protein signatures for multiple transcripts of the same gene.

Our results show at the next nucleotide that order changes, in the vector are rare and orient dramatically to k-mer's of shorter lengths. Further, that the transcripts with the most changes in vector positions are definitive for sequence selections that confer their anti-proliferation effect in transfection experiments.

We anticipate cells approximating these k-mer:protein anti-proliferation signatures will be most useful in the fight against proliferating diseased cells. To test usefulness we will co-culture these cells to precondition Natural Killer cells with conforming and non-conforming receptor-ligand relationship sensitivity. These Natural Killer cells will be tested against HeLa to determine whether recognition and optimal immune response can be triggered.

Stay posted for more updates on our exciting discoveries and computations for minimally manipulated autologous therapy against patient disease.