Machine Learning Discoveries of Bcl-X Synergy in etc-1922159 Treated Colorectal Cancer Cells

1. Introduction

In the unpublished preprint Sinha [1], a frame work of a search engine was developed which can rank combinations of factors (genes/proteins) in a signaling pathway. Such combinations are of import due to the vast search space in which they exist and the difficulty to find them. The search engine facilitates in prioritizing the combinations as ranked biological hypotheses which the biologists might want to test in wet lab, to know if a synergistic combination is prevalent in a signaling pathway, in a direct or indirect manner. Interested readers are advised to go through unpublished preprints Sinha [1] and Sinha [2] for details regarding the search engine and the discoveries mentioned in there.

2. Materials and Methods

2.1. Combinatorial Search Problem and a Possible Solution

The issue of combinatorial search problem and a possible solution has been addressed in Sinha [3] and Sinha [2]. The details of the methodology of this manuscript have been explained in great detail in Sinha [3] & its application in Sinha [2]. Readers are requested to go through the same for gaining deeper insight into the working of the pipeline and its use of published data set generated after administration of ETC-1922159. In order to understand the significance of the solution proposed to the problem of combinatorial search that the biologists face in revealing unknown biological search problem, these works are of importance.

Briefly, from Sinha [2], the pipleline works by computing sensitivity indicies for each of these unique combinations and then vectorising these indices to connote and form discriminative feature vector for each combination. Since each combination is unique, the training and the test data are same. In the training data, the combinations are arranged and ranks from 1 to n are assigned. The ranking algorithm then learns the patterns from these combinations/sensitivity index vectors. Next the learned model is used to rank the test data by generating the ranking score for each of the unique combination. Sorting these shuffled scores of test data leads to prioritization of the combinations. Joachims [4] show an example of applying learned model to training data (same as the test data) in https://www.cs.cornell.edu/people/tj/svm_light/svm_rank.html. Note that these combinations are now ranked and give the biologists a chance to narrow down their focus on crucial biological hypotheses in the form of combinations which the biologists might want to test. Analogous to the webpage search engine, where the click of a button for a few key-words leads to a ranked list of web links, the pipeline uses sensitivity indices as an indicator of the strength of the influence of factors or their combinations, as a criteria to rank the combinations.

3. Results & Discussion

3.1. BCL Related Synergies

3.1.1. Interleukin - BCL cross Family Analysis

Qin et al. [5] observe that IL-6 inhibits starvation-induced autophagy via the STAT3/Bcl-2 signaling pathway. Gabellini et al. [6] observed that interleukin 8 mediates bcl-xL-induced enhancement of human melanoma cell dissemination and angiogenesis in a zebrafish xenograft model. Guruprasath et al. [7] show taht interleukin-4 receptor-targeted delivery of Bcl-xL siRNA sensitizes tumors to chemotherapy and inhibits tumor growth. Maraskovsky et al. [8] indicate that Bcl-2 can rescue T lymphocyte development in interleukin-7 receptor-deficient mice but not in mutant rag-1^−/− mice. Akashi et al. [9] show that Bcl-2 rescues T lymphopoiesis in interleukin-7 receptor-deficient mice. Interleukin-10 increases Bcl-2 expression and survival in primary human CD34+ hematopoietic progenitor cells as shown by Weber-Nordt et al. [10]. Interleukin-7 and interleukin-15 regulate the expression of thebcl-2 and c-myb genes in cutaneous T-cell lymphoma cells as shown by Qin et al. [11]. Bcl-2 is a negative regulator of interleukin-1$\beta$ secretion in murine macrophages in pharmacological-induced apoptosis as shown by Escandell et al. [12]. Alas et al. [13] observe that inhibition of interleukin 10 by rituximab results in down-regulation of bcl-2 and sensitization of B-cell non-Hodgkin’s lymphoma to apoptosis. These findings indicate the synergy between BCL and Interleukin in different pathological cases. In CRC cells treated with ETC-1922159, these were found to be up regulated. Table 1 and Table 2 indicate the rankings of the IL and BCL family.

On the left side is the rankings of IL w.r.t BCL family and the right side, the vice versa. We found IL-1A/1B/17C up regulated w.r.t BCL2L1. These are reflected in rankings of 2482 (laplace) and 1834 (rbf) for IL1A - BCL2L1; 2252 (laplace), 1920 (linear) for IL1B - BCL2L1; and 2481 (laplace), 2410 (linear) and 2512 (rbf) for IL17C - BCL2L1; IL-6ST/17REL were up regulated w.r.t BCL2L2. These are reflected in rankings of 2239 (laplace), 1927 (linear) and 2085 (rbf) for IL6ST - BCL2L2; and 2454 (laplace), 2510 (linear) and 2482 (rbf) for IL17REL - BCL2L2. IL-17REL were up regulated w.r.t BCL2L13. These are reflected in rankings of 2420 (laplace), 2419 (linear) and 2464 (rbf) for IL17REL - BCL2L13; IL-6ST/15RA were up regulated w.r.t BCL3. These are reflected in rankings of 1928 (laplace) and 2344 (rbf) for IL6ST - BCL3; and 2478 (laplace), 1820 (linear) and 2500 (rbf) for IL15RA - BCL3; IL-1RAP/6ST/8/17REL were up regulated w.r.t BCL6. These are reflected in rankings of 2360 (linear) and 1813 (rbf) for IL1RAP - BCL6; 2419 (laplace) and 1962 (rbf) for IL6ST - BCL6; 2363 (laplace) and 2233 (linear) for IL8 - BCL6; and 2253 (laplace) and 2396 (linear) for IL17REL - BCL6; IL-1A/6ST/8/17REL were up regulated w.r.t BCL9L. These are reflected in rankings of 1932 (laplace) and 1942 (linear) for IL1A - BCL9L; 2249 (laplace) and 1960 (linear) for IL6ST - BCL9L; 2197 (linear) and 2162 (rbf) for IL8 - BCL9L; and 2308 (linear) and 1926 (rbf) for IL17REL - BCL9L; IL-6ST/15RA were up regulated w.r.t BCL10. These are reflected in rankings of 2008 (laplace) and 1816 (rbf) for IL6ST - BCL10; and 2064 (linear) and 1789 (rbf) for IL15RA - BCL10;

On the right side is the rankings of BCL w.r.t IL family. We found BCL2L1 up regulated IL-1B/2RG/10RB. These are reflected in rankings of 1838 (laplace) and 2132 (rbf) for IL1B - BCL2L1; 2048 (laplace) and 1949 (rbf) for IL2RG - BCL2L1; and 1965 (linear) and 2024 (rbf) for IL10RB - BCL2L1; BCL2L2 was up regulated IL-1A/1B/1RN/6ST/8/15/17C. These are reflected in rankings of 2407 (laplace), 2362 (linear) and 2464 (rbf) for IL1A - BCL2L2; 1807 (laplace), 2462 (linear) and 2344 (rbf) for IL1B - BCL2L2; 2298 (laplace) and 2092 (rbf) for IL1RN - BCL2L2; 2046 (linear) and 1859 (rbf) for IL6ST - BCL2L2; 1803 (laplace) and 2024 (rbf) for IL8 - BCL2L2; 2474 (laplace), 2142 (linear) and 2416 (rbf) for IL15 - BCL2L2; and 2512 (linear) and 2447 (rbf) for IL17C - BCL2L2; BCL2L13 was up regulated IL-1RAP/1RN/2RG/6ST/8/10RB/15/15RA/17C. These are reflected in rankings of 2450 (linear) and 2510 (rbf) for IL1RAP - BCL2L13; 2503 (laplace) and 2378 (rbf) for IL1RN - BCL2L13; 2483 (laplace) and 2248 (rbf) for IL2RG - BCL2L13; 1899 (laplace), 2473 (linear) and 2046 (rbf) for IL6ST - BCL2L13; 2099 (laplace) and 2294 (rbf) for IL8 - BCL2L13; 2120 (laplace) and 1895 (linear) for IL10RB - BCL2L13; 2515 (laplace), 2160 (linear) and 2420 (rbf) for IL15 - BCL2L13; 1844 (linear) and 2318 (rbf) for IL15RA - BCL2L13; and 2004 (laplace), 2434 (linear) and 2500 (rbf) for IL17C - BCL2L13; BCL3 was up regulated IL-8/10RB. These are reflected in rankings of 2266 (laplace) and 1983 (rbf) for IL8 - BCL3; and 2187 (laplace) and 2170 (rbf) for IL10RB - BCL3; 2298 (laplace); 2423 (linear) and 2294 (rbf) for IL1B - BCL6; 1919 (laplace) and 2301 (linear) for IL1RN - BCL6; 2106 (linear) and 2478 (rbf) for IL2RG - BCL6; 2123 (laplace), 2068 (linear) for IL8 - BCL6; 2084 (laplace), 1791 (linear) and 2203 (rbf) for IL15RA - BCL6; and for 1949 (linear) and 1930 (rbf) for IL17REL - BCL6; BCL10 was up regulated IL-1A/1RAP/1RN/2RG/10RB/15RA. These are reflected in rankings of 2405 (linear) and 1889 (rbf) for IL1A - BCL10; 1929 (laplace) and 2112 (rbf) for IL1RAP - BCL10; 1846 (laplace) and 1823 (linear) for IL1RN - BCL10; 1885 (laplace) and 1803 (linear) for IL2RG - BCL10; 2244 (laplace) and 2150 (linear) for IL10RB - BCL10; and 1810 (laplace) and 1835 (rbf) for IL15RA - BCL10;

Finally, Table 3 shows the derived influences which can be represented graphically, with the following influences - • IL w.r.t BCL with IL-1A/1B/17C $<-$ BCL2L1; IL-6ST/17REL $<-$ BCL2L2; IL-17REL $<-$ BCL2L13; IL-6ST/15RA $<-$ BCL3; IL-1RAP/6ST/8/17REL $<-$ BCL6; IL-1A/6ST/8/17REL $<-$ BCL9L; and IL-6ST/15RA $<-$ BCL10; • BCL w.r.t IL with IL-1B/2RG/10RB $->$ BCL2L1; IL-1A/1B/1RN/6ST/8/15/17C $->$ BCL2L2; IL-1RAP/1RN/2RG/6ST/8/10RB/15/15RA/17C $->$ BCL2L13; IL-8/10RB $->$ BCL3; IL-1B/1RN/2RG/8/15RA/17REL $->$ BCL6; and IL-1A/1RAP/1RN/ 2RG/10RB/15RA $->$ BCL10;

3.1.2. Selenbp1 - BCL cross Family Analysis

Deng et al. [14] study the effects of selenium on lead-induced alterations in A$\beta$ production and Bcl-2 family proteins. Yaming et al. [15] studied the effects of selenium dioxide on apoptosis, Bcl-2 and p53 expression, intracellular reactive oxygen species and calcium level in three human lung cancer cell lines. Activity of selenium on cell proliferation, cytotoxicity, and apoptosis and on the expression of CASP9, BCL-XL and APC in intestinal adenocarcinoma cells has been studied by Mauro et al. [16]. These studies suggest the synergy between BCL and Selenium based genes. In CRC cells treated with ETC-1922159, these were found to be down regulated. Table 4 shows the rankings of BCL family w.r.t to SELENBP1 and vice versa.

On the right side, we found BCL-6B/11A to be up regulated with respect to SELENBP1. These were reflected in the rankings of 182 (laplace), 110 (linear) and 494 (rbf) for SELENBP1 - BCL6B; and 905 (laplace), 931 (linear) and 401 (rbf) for SELENBP1 - BCL11A. On the left side SELENBP1 was up regulated w.r.t BCL-9/11B. These are reflected in rankings of 1568 (linear) and 1738 (rbf) for SELENBP1 - BCL9; and 299 (linear) and 1385 (rbf) for SELENBP1 - BCL11B; Finally, Table 5 shows the derived influences which can be represented graphically, with the following influences - • SELENBP1 w.r.t BCL with SELENBP1 $<-$ BCL-9/11B; and • BCL w.r.t SELENBP1 with SELENBP1 $->$ BCL-6B/11A;

3.1.3. TP53 - BCL cross Family Analysis

The p53-Bcl-2 connection has been studied by Hemann and Lowe [17]. Tomita et al. [18] show wild type p53, but not tumor-derived mutants, bind to Bcl2 via the DNA binding domain and induce mitochondrial permeabilization. Bcl-2 constitutively suppresses p53-dependent apoptosis in colorectal cancer cells as shown by Jiang and Milner [19]. The tissue dependent interactions between p53 and Bcl-2 in vivo has been studied by Li et al. [20]. Synthetic lethality of combined Bcl-2 inhibition and p53 activation in AML has been studied by Pan et al. [21]. Zaidi et al. [22] observe that the chloroquine-induced neuronal cell death is p53 and Bcl-2 family-dependent but caspase-independent. Relationship of p53, bcl-2, and tumor proliferation to clinical drug resistance in non-Hodgkin’s lymphomas has been studied in Wilson et al. [23]. TP53 and BCL family members were found to be up regulated in CRC cells treated with ETC-1922159. Table 6 show rankings of BCL and TP53 family w.r.t to each other.

On the left side, we found BCL2L2 to be up regulated w.r.t TP53-I3/INP2. These are reflected in the rankings of 2423 (laplace), 2377 (linear) and 2452 (rbf) for TP53I3 - BCL2L2; 1827 (linear) and 2035 (rbf) for TP53INP2 - BCL2L2. BCL2L13 to be up regulated w.r.t TP53-INP2. These are reflected in the rankings of 2427 (linear) and 2008 (rbf) for TP53INP2 - BCL2L13; BCL6 to be up regulated w.r.t TP53-I3/INP2. These are reflected in the rankings of 2275 (laplace), 2312 (linear) and 2146 (rbf) for TP53I3 - BCL6; and 2329 (linear) and 2352 (rbf) for TP53INP2 - BCL6; BCL9L to be up regulated w.r.t TP53-BP2. These are reflected in the rankings of 2320 (linear) and 2197 (rbf) for TP53BP2 - BCL9L; BCL10 to be up regulated w.r.t TP53-BP2/INP2. These are reflected in the rankings of 2230 (laplace) and 2418 (linear) for TP53BP2 - BCL10 and 1910 (linear) and 2087 (rbf) for TP53INP2 - BCL10;

On the right side, we found TP53-BP2/I3 to be up regulated w.r.t BCL2L1. These are reflected in the rankings of 1786 (laplace) and 1961 (linear) for TP53BP2 - BCL2L1; 1980 (laplace) and 1752 (linear) for TP53I3 - BCL2L1; TP53-INP1 were up regulated w.r.t BCL3. These are reflected in the rankings for 2259 (linear) and 2043 (rbf) for TP53INP1 - BCL3; TP53-BP2/INP2 were up regulated w.r.t BCL9L. These are reflected in the rankings for 2093 (laplace) and 2217 (linear) for TP53BP2 - BCL9L; and 2222 (laplace) and 1900 (linear) for TP53INP2 - BCL9L;

Finally, Table 7 shows the derived influences which can be represented graphically, with the following influences - • BCL w.r.t TP53 with TP53-I3/INP2 $<-$ BCL2L2; TP53-INP2 $<-$ BCL2L13; TP53-I3/INP2 $<-$ BCL6; TP53-BP2 $<-$ BCL9L; and TP53-BP2/INP2 $<-$ BCL10; • TP53 w.r.t BCL with TP53-BP2/I3 $<-$ BCL2L1; TP53-INP1 $<-$ BCL3 and TP53-BP2/INP2 $<-$ BCL9L.

3.1.4. CASP - BCL cross Family Analysis

Expression of caspase and BCL-2 apoptotic family members in mouse preimplantation embryos have been studied by Exley et al. [24]. Swanton et al. [25] observed that Bcl-2 regulates a caspase-3/caspase-2 apoptotic cascade in cytosolic extracts. Their role in the regulation of the immune response of Caspases, Bcl-2 family proteins and other components of the death machinery has been observed in Pellegrini and Strasser [26]. Moriishi et al. [27] show that Bcl-2 family members do not inhibit apoptosis by binding the caspase activator Apaf-1. In CRC cells treated with ETC-1922159, these families were found to be UP regulated. Table 8 shows rankings of CASP and BCL family.

On the left side, we found BCL2L2 to be up regulated w.r.t CASP-10/16. These are reflected in the rankings of 2043 (linear) and 1809 (rbf) for CASP10 - BCL2L2; and 2263 (laplace) and 1863 (rbf) for CASP16 - BCL2L2; BCL2L13 to be up regulated w.r.t CASP-4/5/16. These are reflected in the rankings of 1873 (laplace) and 2415 (rbf) for CASP4 - BCL2L13; 1962 (laplace), 2514 (linear) and 2493 (rbf) for CASP5 - BCL2L13; and 1762 (laplace), 2492 (linear) and 2166 (rbf) for CASP16 - BCL2L13; BCL3 to be up regulated w.r.t CASP-10. These are reflected in the rankings of 2409 (laplace) and 2011 (linear) for CASP10 - BCL3; BCL6 to be up regulated w.r.t CASP-5/16. These are reflected in the rankings of 1787 (laplace), 2124 (linear) and 2309 (rbf) for CASP5 - BCL6; and 2397 (laplace), 2166 (linear) and 2387 (rbf) for CASP16 - BCL6.

On the right side, we found CASP-5/7 to be up regulated w.r.t BCL2L1. These are reflected in the rankings of 1992 (laplace) and 2053 (linear) for CASP5 - BCL2L1; and 2203 (linear) and 1750 (rbf) for CASP7 - BCL2L1. CASP-4/7 to be up regulated w.r.t BCL2L1. These are reflected in the rankings of 1902 (linear) and 1979 (rbf) for CASP4 - BCL2L13 and 1877 (laplace) and 2216 (rbf) for CASP7 - BCL2L13; CASP-7/16 to be up regulated w.r.t BCL9L. These are reflected in the rankings of 1813 (laplace) and 1980 (rbf) for CASP7 - BCL9L; and 2499 (linear) and 2027 (rbf) for CASP16 - BCL9L; CASP-7 to be up regulated w.r.t BCL10. These are reflected in the rankings of 2489 (laplace) and 1945 (rbf) for CASP7 - BCL10.

3.1.5. MUC - BCL cross Family Analysis

MUC1 and bcl-2 expression in preinvasive lesions and adenosquamous carcinoma of the lung have been studied by Demirag et al. [28]. Sheng et al. [29] report that MUC13 prevents colorectal cancer cell death by promoting two distinct pathways of NF-kB activation, consequently upregulating BCL-X_L. In CRC cells treated with ETC-1922159, family members of BCL and MUC were found up regulated. The search engine assigned high valued numerical ranks to some of the 2^nd order combinations of BCL-MUC family members. Table 10 show the rankings of the members with respect to each other.

On the left side, we found BCL2L1 to be up regulated w.r.t MUC-1/13. These are reflected in the rankings of 2055 (laplace), 2297 (linear) and 1854 (rbf) for MUC1 - BCL2L1; and 1927 (laplace) and 2108 (rbf) for MUC13 - BCL2L1; BCL2L2 was up regulated w.r.t MUC-4/13/17. These are reflected in the rankings of 2506 (linear) and 1988 (rbf) for MUC4 - BCL2L2; 2084 (laplace) and 2402 (linear) for MUC13 - BCL2L2; and 2283 (laplace) and 2212 (linear) for MUC17 - BCL2L2; BCL2L13 was up regulated w.r.t MUC-1/12. These are reflected in the rankings of 2029 (laplace) and 2347 (linear) for MUC1 - BCL2L13; and 2353 (linear) and 1997 (rbf) for MUC12 - BCL2L13; BCL3 was up regulated w.r.t MUC-20. These are reflected in the rankings of 2512 (laplace) and 2440 (rbf) for MUC20 - BCL3; BCL6 was up regulated w.r.t MUC-17. These are reflected in the rankings of 2411(laplace), 2153 (linear) and 1808 (rbf) for MUC17 - BCL6; BCL9L was up regulated w.r.t MUC-17. These are reflected in the rankings of 2101 (laplace) and 2408 (rbf) for MUC20 - BCL9L.

On the right side, we found MUC3A to be up regulated w.r.t BCL2L2. These are reflected in the rankings of 2099 (laplace) and 2397 (rbf) for MUC3A - BCL2L2; MUC3A to be up regulated w.r.t BCL9L. These are reflected in the rankings of 2180 (linear) and 2106 (rbf) for MUC3A - BCL9L;

3.1.6. EXOSC - BCL cross Family Analysis

The exosome complex is involved in the degradation of various kinds of RNA. Recently, Deng et al. [30] observe that Exosome-transmitted LINC00461 promotes multiple myeloma cell proliferation and suppresses apoptosis by modulating microRNA/BCL-2 expression. Xu et al. [31] show that Exosome-derived microRNA-29c induces apoptosis of BIU-87 cells by down regulating BCL-2 and MCL-1. Exosomes were demonstrated to upregulate the expression of Bcl-2 and Cyclin D1 proteins, but reduce the levels of Bax and caspase-3 proteins in these cells in work of Yang et al. [32]. In western blot analysis results showed that exosomes can block the significant reduction of BCL-2, full-length caspase-3 and full-length PARP, while preventing the increase of BAX, cleaved caspase-3 and cleaved PARP induced by VP16, as studied by Wang et al. [33]. These findings point to the definite synergistic role of exosome with BCL family. In CRC cells, both exosome components EXOSC and BCL family members were found to be down regulated, after ETC-1922159 drug treatment. The search engine allocated low numerical valued ranks for many of the EXOSC and BCL combinations which might suggest greater role of EXOSC along with BCL. However, the nature of the mechanism between the two families yet needs to be explored, despite the generated hypothesis of possible synergy.

Table 12 shows rankings of EXOSC and BCL family with respect to each other. Left half of the table shows rankings of EXOSC w.r.t BCL and right half shows the vice versa. On the left, we find EXOSC2 to be down regulated w.r.t BCL-2L12/6B/7A/9/11A/11B. These are shown in the rankings of 723 (laplace), 355 (linear) and 1211 (rbf) for EXOSC2 - BCL2L12; 1092 (laplace), 1033 (linear) and 638 (rbf) for EXOSC2 - BCL6; 1633 (laplace), 1047 (linear) and 317 (rbf) for EXOSC2 - BCL7A; 699 (laplace), 559 (linear) and 425 (rbf) for EXOSC2 - BCL9; 338 (laplace), 319 (linear) and 1598 (rbf) for EXOSC2 - BCL11A; and 1285 (laplace), 1440 (linear) and 812 (rbf) for EXOSC2 - BCL11B; EXOSC3 was found to down regulated w.r.t BCL11B. This is reflected in rankigns of 1677 (laplace), 199 (linear) and 267 (rbf) for EXOSC3 - BCL11B. EXOSC5 was found to be down regulated w.r.t BCL family. These are reflected in the rankings of 498 (laplace), 1342 (linear) and 436 (rbf) for EXOSC5 - BCL2L12; 786 (laplace), 1272 (linear) and 1194 (rbf) for EXOSC5 - BCL6B; 374 (laplace), 1338 (linear) and 874 (rbf) for EXOSC5 - BCL7A; 613 (laplace), 946 (linear) and 772 (rbf) for EXOSC5 - BCL9; 459 (laplace), 90 (linear) and 1034 (rbf) for EXOSC5 - BCL11A; and 1404 (laplace) and 1558 (linear) for EXOSC5 - BCL11B; EXOSC6 was found to be down regulated w.r.t BCL family. These are reflected in rankings of 1676 (laplace), 787 (linear) and 944 (rbf) for EXOSC6 - BCL7A; 1059 (linear) and 1091 (rbf) for EXOSC6 - BCL9; 1677 (laplace) and 1573 (linear) for EXOSC6 - BCL11A; EXOSC7 was found to be down regulated w.r.t BCL family. These are reflected in rankings of 666 (laplace); 98 (linear) and 743 (rbf) EXOSC7 - BCL6B; 1501 (linear) and 1513 (rbf) for EXOSC7 - BCL7A; and 1477 (laplace) and 1217 (rbf) for EXOSC7 - BCL11A; EXOSC8 was found to be down regulated w.r.t BCL family. Thesea reflected in 1175 (laplace), 1504 (linear) and 1743 (rbf) for EXOSC8 - BCL7A; 906 (linear) and 1130 (rbf) EXOSC8 - BCL11A; and 605 (linear) and 374 (rbf) for EXOSC8 - BCL11B; EXOSC9 found to be down regulate w.r.t BCL family. These are reflected in rankings of 1179 (laplace); 1018 (linear) and 687 (rbf) for EXOSC9 - BCL2L12; 437 (laplace), 852 (linear) and 1358 (rbf) EXOSC9 - BCL6B; 821 (laplace), 346 (linear) and 727 (rbf) for EXOSC9 - BCL7A; 1305 (laplace) and 299 (rbf) EXOSC9 - BCL9; 1569 (laplace), 549 (linear) and 1456 (rbf) for EXOSC9 - BCL11B.

On the right, we find BCL-6B/11A/11B to be down regulated w.r.t EXOSC2. These are reflected in the rankings of 202 (laplace), 81 (linear) and 194 (rbf) for EXOSC2 - BCL6B; 574 (laplace), 834 (linear) and 1055 (rbf) for EXOSC2 - BCL11A; and 1368 (laplace), 1353 (linear) and 1455 (rbf) for EXOSC2 - BCL11B. BCL-6B/7A/11A was found to be down regulated w.r.t EXOSC3. These are reflected in rankings of 571 (laplace), 335 (linear) and 307 (rbf) for EXOSC3 - BCL6B; 1739 (laplace) and 1700 (rbf) for EXOSC3 - BCL7A; and 1018 (laplace), 1345 (linear) and 483 (rbf) for EXOSC3 - BCL11A; BCL-6B/11A/11B was found to be down regulated w.r.t EXOSC5. These were reflected in rankings of 571 (laplace), 335 (linear) and 307 (rbf) for EXOSC5 - BCL6B; 756 (laplace), 389 (linear) and 1183 (rbf) for EXOSC5 - BCL11A; and 1368 (laplace), 1353 (linear) and 1455 (rbf) for EXOSC5 - BCL11B. BCL-9 was found to be down regulated w.r.t EXOSC6. These are reflected in rankigns of 851 (linear) and 1564 (rbf) for EXOSC6 - BCL9. BCL-2L12/7A/9/11A/11B was found to be down regulated w.r.t EXOSC7. These are reflected in rankings of 1551 (linear) and 1099 (rbf) for EXOSC7 - BCL2L12; 1282 (laplace), 831 (linear) and 1218 (rbf) for EXOSC7 - BCL7A; 1234 (linear) and 328 (rbf) for EXOSC7 - BCL9; 520 (laplace), 117 (linear) and 686 (rbf) for EXOSC7 - BCL11A; and 1529 (laplace) and 1418 (rbf) for EXOSC7 - BCL11B; BCL-6B/11A was found to be down regulated with EXOSC8. These are reflected in rankings of 190 (laplace), 1630 (linear) and 472 (rbf) for EXOSC8 - BCL6B; and 944 (laplace) and 532 (rbf) for EXOSC8 - BCL11A. Finally, BCL-6B/9/11A/11B was found to be down regulated with EXOSC9. These are reflected in rankings of 634 (laplace), 304 (linear) and 146 (rbf) for EXOSC9 - BCL6B; 1197 (laplace) and 1279 (rbf) for EXOSC9 - BCL9; 481 (laplace), 441 (linear) and 1372 (rbf) for EXOSC9 - BCL11A; and 1454 (linear) and 133 (rbf) for EXOSC9 - BCL11B.

Table 13 shows the derived influences which can be represented graphically, with the following influences - • EXOSC w.r.t BCL with EXOSC2 $<-$ BCL-2L12/6B/7A/9/11A/11B; EXOSC3 $<-$ BCL-11B; EXOSC5 $<-$ BCL-2L12/6B/7A/9/11A/11B; EXOSC6 $<-$ BCL-7A/9/11A; EXOSC7 $<-$ BCL-6B/7A/11A; EXOSC8 $<-$ BCL-7A/11A/11B and EXOSC9 $<-$ BCL-2L12/6B/7A/9/11B; and • BCL w.r.t EXOSC with EXOSC2 $->$ BCL-6B/11A/11B; EXOSC3 $->$ BCL-6B/7A/11A; EXOSC5 $->$ BCL-6B/11A/11B; EXOSC6 -> BCL-2L12/9; EXOSC7 $->$ BCL-2L12/7A/9/11A/11B; EXOSC8 $->$ BCL-6B/11A and EXOSC9 $->$ BCL-6B/9/11A/11B.